Socket
开发者优先的软件供应链安全已有真实客户验证,但 $1B 估值下经济质量仍不透明
Socket 在软件供应链安全里已经跑出真实产品市场匹配:AI / 开发者客户证据强、席位定价透明、行为加可达性技术栈差异化;但 ARR、留存、利润率、烧钱速度和股权条款仍未披露,仅凭公开证据看,2026 年 5 月 $1 billion Series C 仍略显偏高。
封面要素
公司概况
Socket 是一家位于 San Francisco 的软件供应链安全公司,由创始人兼 CEO Feross Aboukhadijeh 领导。Socket 的 About 页面称公司创立于 2021 年,但若干 2026 年融资材料称为 2020 年。公司销售开发者优先的平台, 覆盖 GitHub、CLI、VS Code、Firewall、API 和 SDK 工作流,用于拦截恶意软件包,在合并或安装前暴露依赖风险, 现在还借助源自 Coana 的可达性分析分流 CVE。2026 年 5 月公开材料称,Socket 保护超过 27,000 个组织、 1.5 million 个代码库和每月 11.6 million 次提交,客户包括 Anthropic、xAI、Replit、Cursor、Figma、Vercel、 Gusto、Mercado Libre 和 Cribl。2026 年 5 月 Series C 将 Socket 估值推至 $1 billion,累计已披露融资达到 $125 million。公开披露仍未给出 ARR、毛利率、NRR、烧钱速度或付费席位转化。
- 成立时间
- 2021-01-01
- 创始人
- Feross Aboukhadijeh
- 创立地点
- San Francisco, California (public materials consistently place Socket there; separate founding-location disclosure not found)
- 总部
- San Francisco, California, USA
- 产品
- 面向开源依赖的开发者工作流安全平台,覆盖 GitHub App、CLI、VS Code 扩展、Socket Firewall、REST API 和 SDK。 核心产品把基于行为的软件包与维护者分析,同安装时拦截以及源自 Coana 的可达性分析结合起来;企业层级增加合规集成、 SBOM 工作流、SSO/SAML、审计日志、自定义策略和更广的生态覆盖。
- 客户
- AI 原生、云和安全意识较强的软件团队,希望把依赖防护嵌入现有 GitHub 中心工作流;典型买方是 CISO、 安全工程或平台安全负责人,日常用户则是开发者和平台工程师。
- 商业模式
- 按活跃开发者计价的订阅 SaaS:开源使用免费,Team 每名开发者每月 $25,Business 每名开发者每月 $50, Enterprise 自定义。Socket 靠自助式 GitHub 部署切入,再扩展到企业控制、可达性、Firewall、合规工作流和市场采购。
- 阶段
- Series C / late-stage private (May 2026 $1B round)
- 融资情况
- Socket 于 2024 年 10 月完成 $40 million Series B,并于 2026 年 5 月完成 $60 million Series C, 累计已披露融资达到 $125 million。Series C 由 Thrive Capital 领投,Andreessen Horowitz、Abstract Ventures 和 Capital One Ventures 参投;公司公开材料还强调 Elad Gil、Bret Taylor 以及 Patrick 和 John Collison 的支持。
执行摘要
主要优势
- 行为优先的依赖安全叠加 Coana 可达性,让 Socket 相比只看 CVE 的工具有了差异化工作流叙事,并可能显著降低告警噪音。
- 以当前阶段看,客户证据异常强,公开引用覆盖 Anthropic、xAI、Replit、Cursor、Vercel、Figma、Gusto、Mercado Libre、Cedar、JumpCloud、Render 和 JupiterOne。
- 分发路径对开发者友好:GitHub 原生落地、免费开源入口、透明 Team / Business 定价,以及从自助到企业的扩张路径。
- 产品节奏很快,2025-2026 年扩展到 Firewall、可达性、PHP/Composer、OpenVSX、Jira、Data Exports,以及 AI 工具或 MCP 覆盖。
- 高质量投资人和轻资本 SaaS 交付模型,让 Socket 能继续扩张,而不必承担硬件或库存风险。
主要风险
- 公开证据仍缺 ARR、付费客户数、NRR、毛利率、烧钱速度和股权结构优先权条款,$1B 融资轮很难承保。
- 如果 Socket 不能持续守住明显更好的信号质量和工作流契合度,GitHub 原生打包以及 Snyk 等更宽 AppSec 平台会挤压变现。
- 公开证据在 GitHub/npm 和 JavaScript 重度环境中最强;离开文档最充分的生态后,文档和外部测试显示证据更弱且不均衡。
- 告警噪音是真风险:Socket 自己提醒 AI 辅助检测可能产生误报,公开 issue 或评价也显示部分良性包和覆盖范围投诉。
- 法务和合规界面落后于产品界面,包括隐私政策最后更新于 2022 年,以及合同责任或赔偿条款公开可见度有限。
未决问题
- 当前 ARR 或 GAAP 收入,以及付费与免费开发者转化;$1B 估值对变现 ARR 是否已经达到数千万美元中段高度敏感。
- 净收入留存、毛利率、烧钱效率和现金跑道;没有公开来源给出检验软件质量经济性的核心耐久指标。
- Series C 的完全摊薄股权结构、清算优先权和任何老股流动性条款。
- Coana 收购价格、整合成本,以及收购后可衡量的增购或留存影响。
- 跨生态产品一致性,以及最清晰的 GitHub/npm 或 JavaScript 重度部署之外的大企业证据。
- 巨大的受保护组织数和仓库数,是否能转化成持久的付费客户集中度和续约质量。
目录
01公司概况
1.1 身份、产品逻辑与当前规模
Socket 的公开材料一贯把公司定位为开发者优先的软件供应链安全平台:恶意或高风险开源依赖进入生产前就要被拦下。 核心产品逻辑不是先查数据库,而是先看行为。Socket 称会实时分析依赖行为,再把结果推到 GitHub、CLI、文档和安装时 Firewall 工作流里,而不是等 CVE 披露后再入库。文档和定价材料把商业定位说得更清楚:开源项目有免费路径, 企业买方为策略深度和支持付费;Socket 还称源代码本身留在本地,只把依赖元数据传到上游。 身份层基本清楚,但并不完全干净。当前官方和独立材料都把 Socket 放在 San Francisco,但公开来源对创立年份不一致: 官方 About 页面写 2021,若干 2026 年融资材料写 2020。这个差异不改变经营故事,但提醒我们,公司公开时间线仍需要文件层面的校准。 有充分支撑的是规模。到 2026 年 5 月,Socket 称其保护超过 27,000 个组织、1.5 million 个代码库、每月 11.6 million 次提交, 并且每周拦截超过 10,000 次攻击。[CO001, CO002, CO003, CO004, CO005, CO009]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 缺口 / 备注 |
|---|---|---|---|---|
| 创立年份 | 官方 About 页面为 2021;多份 2026 融资材料为 2020 | 2026 视角 | 中 | 公开来源不一致;需核验注册记录和上线时间线 |
| 总部 | San Francisco, California | 2024-2026 | 高 | 城市已得到佐证;公开材料没有给出标准运营办公室地址 |
| 核心产品 | 面向开发者、保护开源依赖的软件供应链安全 | 2026 | 高 | 行为检测定位在官方材料中保持一致 |
| 投后估值 | $1B | 2026-05-20 | 高 | 基于 Series C 披露 |
| 累计融资 | $125M | 2026-05-20 | 高 | 基于 Series C 披露 |
| 受保护组织 | 27,000+ | 2026-05 | 高 | 公司报告的运营指标,未经审计 |
| 受保护仓库 | 1.5M | 2026-05 | 高 | 公司报告的运营指标,未经审计 |
| 每月受保护 commits | 11.6M+ | 2026-05 | 高 | 公司报告的运营指标,未经审计 |
| 每周拦截攻击 | 10,000+ | 2026-05 | 高 | 公司报告的运营指标,未经审计 |
| 团队规模 | 100+ people | 2026-05 | 中 | 当前员工数是方向性指标,不是精确发薪人数 |
| 收入 / ARR | 已审查来源未公开披露 | 2026-05-24 | 低 | 需要管理层 KPI 包或董事会材料来评估商业效率 |
快照结合了官方运营指标和公开第三方佐证;规模数字由公司报告,收入仍未披露。
[CO003, CO004, CO005, CO011, CO012, CO013]AI 驱动的编程需求、依赖行为分析、安装时拦截、客户证据和创始人集中度,如何串起 Socket 当下的公司逻辑。
[CO001, CO002, CO018, CO023, CO027, CO030]1.2 创始人杠杆、技术班底与客户验证
领导结构明显偏创始人中心。Feross Aboukhadijeh 仍是 About 页面、融资帖子、技术叙事和招聘材料里的公开主角; 他作为高产开源维护者、Node.js 治理参与者和 Stanford 讲师的经历,也正好贴合 Socket 要解决的问题。 这种创始人与市场的匹配是真实战略资产:客户和投资人反复把开发者可信度列为 Socket 能替代传统 SCA 产品的原因。 代价是集中度。已审阅的公开材料把 Feross 展示得很清楚,却没有对成熟高管班底、正式董事会结构或投资人治理权给出同等细节。 客户证据比治理披露更强。2024 年和 2026 年官方材料反复点名 Anthropic、xAI、Replit、Cursor、Figma、Vercel、 Gusto、Mercado Libre 和 Cribl;文档和更早的客户引语又分别印证 Brave、MetaMask 以及更广 JavaScript 生态开源项目的使用。 这个模式说明,Socket 已经最能打动高速推进、最暴露于依赖风险的 AI 原生和开发者基础设施团队。Coana 和 Secure Annex 收购在组织上也重要: 它们等于把专业技术负责人带进公司,并把覆盖范围从经典软件包扫描向外拓宽。[CO006, CO007, CO008, CO018, CO019, CO029]
| 人物 | 角色 | 背景 | 创始人-市场匹配 / 职能覆盖 | 关键人物依赖 |
|---|---|---|---|---|
| Feross Aboukhadijeh | 创始人兼 CEO | WebTorrent 和 StandardJS 创作者;Node.js 治理参与者;Stanford 讲师 | 直接开源可信度和开发者同理心,与软件供应链安全高度贴合 | 高 |
| Anders Søndergaard(Coana 联合创始人) | Coana 联合创始人兼前 CEO;通过收购加入 Socket | 基于 Aarhus University 研究,搭建可达性分析和静态分析工具 | 增强精准 CVE 分诊深度,帮助缓解误报疲劳 | 中 |
| John Tuckner | Secure Annex 创始人;通过 2026 收购加入 | 扩展安全研究员和独立创始人 | 将 Socket 覆盖范围从包扩到浏览器、IDE 和 AI 工具界面 | 中 |
| 公开治理可见度 | 创始人清晰;更广董事会和高管披露不清 | 公开材料强调投资人和收购,多于正式治理图 | 在董事会席位、委员会和投资人控制条款上留下尽调缺口 | 中 |
这是面向公开信息的部分团队图谱,不是完整组织架构;公开来源高度强调创始人和收购带来的技术补强。
[CO006, CO007, CO029, CO030, CO038, CO044]1.3 融资形成、投资人地图与里程碑执行
Socket 的资本故事现在是最清楚的外部验证信号。2024 年 10 月 Series B 融入 $40 million,由 Abstract Ventures 领投, 累计融资增至 $65 million;2026 年 5 月 Series C 又以 $1 billion 估值新增 $60 million,累计融资升至 $125 million。 C 轮由 Thrive Capital 领投,a16z、Abstract Ventures 和 Capital One Ventures 参投,让公司的股权结构叙事同时包含风投品牌、 安全可信度和企业分发可选项。公司自己的投资人页面也强化了这个故事:支持者来自安全运营者、开源领袖和高知名度技术创始人。 两轮之间的里程碑不只是融资通胀。2024 年 10 月,Socket 公开报告 7,500 个受保护组织和 300,000 个 GitHub 代码库。 到 2025 年 4 月,Coana 收购与可达性分析主张绑定,指向传统 SCA 的误报疲劳。到 2025 年 9 月,Socket 推出 Socket Firewall Free,把防护前移到安装点。到 2026 年 5 月,公司引用 27,000+ 个组织、1.5 million 个代码库、 每月 11.6 million 次受保护提交以及超过 100 名员工。即便收入等关键商业指标仍未披露,这仍是快速执行,而不只是纯风投叙事。[CO014, CO015, CO016, CO017, CO020, CO021]
| 利益相关方 | 角色 | 控制 / 经济重要性 | 证据 | 尽调问题 |
|---|---|---|---|---|
| Feross Aboukhadijeh | 创始人兼 CEO | 核心经营者和叙事所有者;可能对产品和招聘有关键影响 | About、融资和招聘页面都以创始人为中心 | 确认投票控制权、董事会角色和继任深度 |
| Thrive Capital | Series C 领投方 | 当前 $1B 估值跃升的主要资金提供方 | 领投 May 2026 Series C | 澄清董事会席位、pro rata 权利和增长预期 |
| Abstract Ventures | Series B 领投方和持续投资人 | 支持 2024 拐点轮,并继续留在 2026 投资团 | 领投 Series B;再次参与 Series C | 核验持股比例和 follow-on reserve 策略 |
| Andreessen Horowitz(a16z) | 重复投资人和市场验证者 | 同时出现在 Series B 和 Series C 叙事中;安全网络信号很强 | 在公司材料中被引用,并列入两轮融资 | 判断 a16z 是否拥有正式治理权,还是主要提供信号价值 |
| Capital One Ventures | Series C 新战略投资人 | 可能带来企业渠道和受监管行业触达 | 被列为 Series C 新参与方 | 评估是否存在商业 GTM 协议,还是只有财务持股 |
| 旗舰客户队列 | 参考客户和需求验证者 | Anthropic、Replit、Vercel、Figma、xAI 等支撑市场可信度 | 在 2024-2026 材料中反复具名 | 量化 ARR 集中度、部署广度和续约行为 |
图谱聚焦公开可读的利益相关方;持股比例、清算优先权、董事会席位和客户集中度在此未披露。
[CO018, CO020, CO022, CO023, CO024, CO038]| 日期 | 事件 | 类型 | 金额 / 估值 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2021 | 官方 About 页面反映 Socket 创立 | 创立 | 官方材料称 2021 创立;部分 2026 来源称 2020 | Feross Aboukhadijeh | 起始年份需要文档层面对齐,才能成为标准口径 |
| 2024-10-22 | Series B 公布 | 融资 | $40M;累计融资 $65M | Abstract Ventures、a16z、Elad Gil 和天使投资人 | 确立机构背书,并为早期采用之后的增长提供资金 |
| 2024-10-22 | Series B 公告发布客户背书 | 合作 | Anthropic、Replit、Figma、Vercel 等被引用 | 客户高管和安全负责人 | 释放出 AI-native 和高变化工程团队中的产品可信度信号 |
| 2024-10-22 | Series B 后公开规模节点 | 规模 | 7,500+ 个组织;300,000 个 GitHub 仓库 | Socket | 显示 unicorn 轮之前已有早期企业牵引力 |
| 2025-04-25 | Coana 收购公布 | 产品 | 加入可达性分析;声称误报降低 80% | Socket 和 Coana | 相比传统 SCA 工作流,提升精准度和 CVE 优先级排序 |
| 2025-04-25 | Coana 团队加入 Socket | 治理 | 收购式招聘创始团队和研究团队 | Anders Søndergaard、Anders Møller、Martin Torp 与 Benjamin Barslev | 加深技术班底,并把产品可信度推向精准分析 |
| 2025-09-30 | Socket Firewall Free 上线 | 产品 | 面向 JS/TS、Python 和 Rust package managers 的免费安装时拦截 | Socket | 把防护前移到安装点,而不是下载后的扫描 |
| 2026-03-20 | Socket CLI issue tracker 上出现公开误报投诉 | 反向 | 良性 textlint package 因 URL strings 风险被标记 | GitHub 用户 h13 | 说明启发式和 AI 辅助检测背后的信任负担 |
| 2026-05-20 | Series C 公布 | 融资 | $60M,估值 $1B;累计融资 $125M | Thrive Capital、a16z、Abstract Ventures 与 Capital One Ventures | 标志 Socket 迈入 unicorn 阶段,并获得更广平台扩张资金 |
| 2026-05-20 | 最新公开规模节点 | 规模 | 27,000+ 个组织;1.5M 个仓库;每月 11.6M 次提交;100+ 人团队 | Socket | 展示 Series B 后足迹和运营规模大幅增长 |
这是第 1 章的公开记录时间线;它有意保持不完整,因为部分上线和注册日期在已审查来源中披露并不一致。
[CO014, CO020, CO021, CO022, CO024, CO027]按时间梳理 Socket 的公开拐点:成立、融资、产品扩展、收购,以及首个可见精准度风险信号。
Secure Annex 条目以年份精度展示,因为抓取到的公告可读文本没有给出规范发布日期。
[CO004, CO020, CO027, CO030, CO031, CO036]1.4 执行风险、披露缺口与尽调优先级
第一章的主要风险,并不是公开可见的诉讼或融资短缺,而是精度、披露和信任。Socket 的卖点取决于实时、AI 辅助的安全判断: 它要早到足以拦截坏软件包,又不能变成另一个噪声扫描器。公司自己承认这种权衡:Firewall Free 在纯 AI 信号上只做警告, 因为可能出现误报。关于 Firewall 发布的独立报道重复了同样限制;2026 年 3 月的一条 GitHub issue 还显示,有用户把一个良性软件包告警质疑为误报。 这些证据不否定产品,但确实强调,检测质量是用户信任和留存的核心。 第二个风险是信息不对称。公开材料在客户、投资人、收购和产品发布上信息丰富,但 ARR、客户集中度、董事会组成、债务和任何老股流动性活动都很薄。 这意味着后续尽调章节可以有信心复用这里的身份、客户和资本事实,但不能假设经济质量或治理也有同样能见度。眼下的关键要求是: 校准 2020/2021 创立年份冲突,拿到干净的董事会和股权结构表,量化 ARR 与净留存,并测试 Coana 时代的精度改进是否在生产账号里可衡量地降低告警疲劳。[CO004, CO033, CO035, CO036, CO037, CO038]
Socket 当前成熟度、牵引力和风险姿态的高层评分视图;绝对经济指标未公开处,用定性分数替代。
分数是基于引用主张推导出的序数综合值,不是公司发布的 KPI。
[CO018, CO022, CO024, CO033, CO035, CO036]1.5 图表要点
02市场分析
2.1 市场边界、纳入支出与现状替代品
Socket 应被界定为软件供应链安全公司,而不是整个应用安全的替身。直接采购问题是判断第三方软件包或更新是否足够安全, 能否进入代码库、CI 流水线或生产发布。Socket 自己的产品界面强调脆弱和恶意依赖、PR 门禁,以及拼写仿冒、安装脚本、 混淆、shell 访问、网络访问和环境变量访问等风险信号。这意味着纳入支出包括依赖审查、恶意软件包检测、支持 SBOM 的资产清单、 安全公告监控,以及同软件交付绑定的策略或分流工作流。排除支出则是多数独立 SAST、DAST、API 测试和一般云安全, 除非买方重开更大的 AppSec 平台合同。替代品集合异常深。Dependabot、npm audit、OSV、Dependency-Check、 Dependency-Track 和 Renovate 为更新、CVE 匹配或资产清单提供低成本或免费基线;GitHub、GitLab、Snyk 和 Black Duck 则把依赖控制打包进更广平台。尽调时,市场边界应当是依赖准入和软件供应链控制这个经常性工作流,而不是所有贴着 AppSec 标签的支出。[CM001, CM002, CM003, CM004, CM005, CM006]
| 细分 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 相关性 |
|---|---|---|---|---|
| 直接依赖安全 / SCA | 依赖准入控制、版本更新审查、CVE 匹配、许可证和元数据审查、PR 卡点 | 大多数独立 SAST、DAST、API 测试和通用云安全 | 工程平台、AppSec 或共享工程 / 安全预算 | Socket 的核心直接品类 |
| 恶意包和行为检测 | 面向包和更新的 typosquat、安装脚本、混淆、网络、shell 和 credential-risk 检测 | 通用代码质量工具或仅运行时防护 | 安全工程和开发者平台负责人 | 超越 CVE-only 扫描的主要溢价切入口 |
| SBOM / 清单 / 策略工作流 | SBOM 生成或摄取、安全公告刷新、清单、策略例外、VEX 或 VDR 邻近工作流 | 不具备 package intelligence 的通用 GRC 工具 | 安全、合规、采购和平台团队 | 监管驱动的邻近场景,仍能服务 Socket |
| 内置 repo-host 覆盖 | 嵌入 SCM 和 CI 工作流的 GitHub 或 GitLab 依赖安全功能 | 无关的源代码控制协作支出 | 现有 GitHub 或 GitLab 平台所有者 | 强替代品和分发压力 |
| 开源和免费工具 | npm audit、OSV、Dependency-Check、Dependency-Track、Renovate 及类似社区工具 | 高阶托管服务或企业支持合同 | 维护者、开发者和成本敏感团队 | 为基础扫描和清单设定价格底线 |
| 更广 AppSec 平台邻近 | 与 SAST、DAST、密钥扫描和更广开发者安全套件一起销售的 SCA | 纯网络或端点安全预算 | CISO、AppSec 或企业平台买方 | 可作为 TAM 上限,但不是干净的直接 SAM |
经济上有意义的边界,是开发者工作流里的依赖准入和软件供应链控制;更广 AppSec 是邻近市场,开源和内置工具则定义基线替代方案。
[CM001, CM002, CM003, CM004, CM005, CM006]合适的市场框架,应从宽泛 AppSec 邻近市场收窄到更小的依赖安全切口;这个切口由付费开发者工作流覆盖定义。
公开数据下,要给 Socket 能切入的市场划出可辩护边界,只能把品类规模估计和基于定价的可服务市场(SAM)逻辑放在一起。
[CM016, CM017, CM019, CM020, CM021, CM043]2.2 规模测算:直接供应链安全、更广 AppSec 与可服务市场(SAM)
公开市场规模需要按区间处理,而不是只看单一总可用市场(TAM)标题。我们找到最干净的直接品类口径来自 Verified Market Reports: 软件供应链安全 2025 年为 USD 1.2 billion,到 2034 年增至 USD 4.5 billion。更广应用安全估算大一个数量级: Mordor 给出 2026 年 USD 14.83 billion,Fortune 给出 2026 年 USD 14.86 billion。这些数字可作为相邻天花板, 因为更广 AppSec 预算有时会吸收依赖安全,但它们太宽,不能称为 Socket 的直接 TAM。另一个极端是 Mordor 的 SCA 页面声称 2026 年市场为 USD 430.12 billion;相对邻近 AppSec 估算过大,应视为品类膨胀警示,而不是估值锚。实际测算应分层: 直接软件供应链安全作为地板,更广 AppSec 作为相邻天花板,可服务市场(SAM)则由活跃开发者或提交者界定——其组织愿意为依赖准入控制、 SBOM 工作流和恶意软件包分流持续付费。[CM012, CM013, CM014, CM015, CM016, CM017]
| 视角 | 发布方 | 年份 / 期间 | 地理 | 数值 | CAGR | 方法 | 置信度 | 局限 |
|---|---|---|---|---|---|---|---|---|
| 直接软件供应链安全 | Verified Market Reports 市场报告 | 2025-2034 | 全球 | 2025 为 USD 1.2B,2034 达 USD 4.5B | 16.5% | 直接软件供应链安全市场快照 | 中 | 最接近的直接市场视角,但底层方法仍不透明 |
| 更广应用安全邻近 | Mordor Intelligence | 2026-2031 | 全球 | 2026 为 USD 14.83B,2031 达 USD 28.11B | 13.64% | 更广应用安全市场预测 | 中 | 覆盖范围远超依赖安全 |
| 更广应用安全邻近 | Fortune Business Insights | 2026-2034 | 全球 | 2026 为 USD 14.86B,2034 达 USD 43.28B | 14.3% | 更广应用安全市场预测 | 中 | 同样有邻近市场问题;不是 Socket 的直接市场 |
| 过宽 SCA 上限 | Mordor Intelligence | 2026-2031 | 全球 | 2026 为 USD 430.12B,2031 达 USD 981.62B | 17.95% | 软件成分分析页面 | 低 | 相比邻近 AppSec 估算明显过宽,不适合作为直接 TAM 锚点 |
| 变现视角 | GitHub | 当前 | 全球 | 每名活跃 committer 每月 USD 19 to USD 30 | n/a | GitHub 官方附加组件定价 | 中 | 定价是工作流代理,不是市场规模估算 |
| 变现视角 | Snyk | 当前 | 全球 | Free、Team、Ignite(<50 developers)、Enterprise;200 / 1000 / 自定义 SCA 测试容量 | n/a | Snyk Open Source 官方套餐包装 | 中 | 配额包装,不是总市场需求 |
| 变现视角 | GitLab | 当前 | 全球 | 依赖扫描打包在 Ultimate enterprise tier | n/a | 官方平台层级打包 | 中 | 未披露独立依赖安全价格 |
直接市场规模是下限,广义 AppSec 是相邻上限,每开发者或每提交者定价则用来搭出与 Socket 相关的 SAM。Mordor 的 SCA 数据保留为矛盾证据,不作为估值锚点。
[CM012, CM013, CM014, CM015, CM016, CM017]公开品类估计的跨度很大:下限是直接的软件供应链安全市场,上限则扩到更宽的 AppSec 与 SCA 口径。
数值单位为十亿美元。直接口径是 2025 年下限,AppSec 口径是 2026 年邻近市场估计;SCA 口径过宽,图中保留为矛盾项,而不是把它抹平。
[CM012, CM013, CM015, CM016, CM043]2.3 买方、用户、付款方分层与采用路径
初始用户通常是开发者、平台工程师或构建负责人;一旦治理变重要,最终付款方往往会切换。Socket 围绕过去 90 天向被扫描代码库提交代码的开发者计价; GitHub 对 Code Security 和 Secret Protection 按活跃提交者收费;GitLab 把更强的依赖安全放入 Ultimate; Snyk 则从免费个人使用一路爬升到团队和企业计划。这些包装方式指向一种常见市场结构:工具先自下而上落进代码库工作流, 等 AppSec、平台或合规负责人需要集中策略、报告和支持时再变现。免费和开源工具对采用路径至关重要,因为它们设定了团队不付费也能拿到什么能力的基线。 Dependabot、npm audit、OSV、Dependency-Check、Dependency-Track 和 Renovate 能处理基本更新自动化、CVE 可见性或 SBOM 清单。 买方只有在相信内置选项漏掉恶意行为、告警噪声过高,或无法满足采购和审计要求时才会升级。对 Socket 而言,最适配的细分市场是: 组织已经感到免费或打包工具的边界,但仍想要比完整 AppSec 套件更轻、更准的东西。[CM017, CM018, CM019, CM020, CM021, CM022]
| 细分市场 | 买方 | 用户 | 付费方 | 工作流 | 预算归属 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 开源维护者和小团队 | 维护者或工程负责人 | 开发者 | 同一团队或无人付费 | 更新依赖、审查 PR、运行免费扫描 | 工程团队或无明确归属 | 零散 CVE 可见性不够 |
| 成长期 SaaS 工程团队 | 工程经理或平台负责人 | 开发者和平台工程师 | 工程预算 | 在 PR 里卡住依赖变更,并监控新公告 | 工程平台 | 免费工具噪音过大,或第一次恶意包警报 |
| 中央 AppSec / 安全工程 | AppSec 负责人 | 开发者和安全分析师 | 安全预算 | 设定策略、批准例外,并在各 repo 统一分流 | AppSec 或 CISO 团队预算 | 需要集中报告、支持和可审计性 |
| 平台 / DevOps 与构建负责人 | 平台或 SRE 负责人 | CI/CD 运维人员和开发者 | 平台预算 | 保护构建流水线、runner 和包解析路径 | 平台工程预算,安全团队背书 | 担心 CI 密钥、传递性风险或供应链事件 |
| 受监管企业和采购主导买方 | CISO、采购或合规负责人 | 开发者、AppSec 和审计人员 | 安全 / 合规预算 | 产出 SBOM 证据、持续重扫和生命周期控制 | 安全、风险或合规 | EO 14028、CRA、客户问卷或审计 |
| 优先打包采购的企业平台买方 | 工程 VP、CIO 或 CISO | 开发者 | 平台 / 安全共用预算 | 默认选 GitHub、GitLab 或更宽的 AppSec 套件 | 现有平台负责人 | 除非最佳单点工具明显更强,否则偏好供应商整合 |
工具从单一 repo 走向全组织治理后,买方、用户和付费方通常会分离。内置工具和开源工具占住最早阶段;付费供应商若能明显降噪或补上合规证据,才会在后期胜出。
[CM017, CM018, CM019, CM020, CM021, CM022]这张图拆出依赖安全采购中谁先启动决策、后续谁掌握预算,以及打包替代方案在哪个环节拦截购买。
[CM017, CM019, CM020, CM021, CM024, CM032]采用从开发者工作流中的依赖变更开始,在事故和监管压力下变硬,最后撞上套件打包压力。
[CM028, CM032, CM035, CM040, CM042, CM043]2.4 增长驱动、采用约束与估值意义
这个品类有真实结构性顺风。EO 14028 要求 NIST 推进软件供应链安全;NIST 的 SSDF 明确给采购方和使用方一套可用于采购的框架; CISA 把 SBOM 称为软件供应链风险管理的关键构件;EU Cyber Resilience Act 则创造了生命周期网络安全义务, 并从 2026 年 9 月开始启动报告要求。安全事件持续强化政策压力。XZ 证明,上游软件包可以被植入后门,并触达 ssh 生产路径; Apache 的 Log4j 安全页面仍证明,高严重性事件之后,传递依赖响应可能拖很久。威胁遥测仍然强烈:Sonatype 描述了工业化代码库滥用和开发者或 CI 环境中的密钥外泄, Veracode 则显示恶意 URL、混淆和拼写仿冒激增。约束同样真实。买方面对误报疲劳、技能短缺和工具蔓延; GitHub、GitLab、Snyk 及其他打包产品又压缩独立定价。因此,Socket 的上行空间不那么取决于市场是否增长, 更取决于其行为优先检测能否比免费和打包替代品带来显著更好的结果。[CM028, CM029, CM030, CM031, CM032, CM033]
| 驱动因素 / 约束 | 方向 | 时点 | 含义 | 尽调追问 |
|---|---|---|---|---|
| EO 14028、NIST SSDF、CISA SBOM 指引和 EU CRA 义务 | 正向 | 当前 | 供应链安全证据进入采购语境,买方紧迫感不再只来自工程便利性 | 追问哪些受监管行业或企业客户因为 SBOM 或安全开发要求,把 Socket 拉进评审 |
| 对开源的依赖和更大的传递性依赖图 | 正向 | 当前 | 底层问题面长期存在,依赖控制从可选项变成必选项 | 量化哪些场景里依赖树太大、人工审查跑不动,因而 Socket 胜出 |
| XZ 和 Log4Shell 带来的事件记忆 | 正向 | 当前 | 让管理层持续盯住上游和传递性组件风险 | 复盘哪些 pipeline 来源是在高知名度依赖事件后专门打开的 |
| AI 生成代码和不断上升的传递性依赖量 | 正向 | 当前 | 审查量上升,买方更容易转向自动分流 | 验证 AI 辅助编码是否显著改变 pipeline 中的扫描量或升级紧迫性 |
| 内置和开源替代品 | 负向 | 当前 | 入门价格下限被压低,付费转化被推迟 | 衡量 Socket 多大比例是在替代免费工具,还是与免费工具并存 |
| 误报疲劳和噪音告警队列 | 负向 | 当前 | 买方对通用扫描更怀疑,新工具需要更强证明 | 要求提供证据,说明 Socket 相比只看 CVE 的工具能显著降低分流负担 |
| 技能缺口和总拥有成本 | 负向 | 当前 | 成本敏感团队铺开更慢,部分买方会转向托管或打包选项 | 追问上线时间、服务依赖,以及停滞交易中的买方画像 |
| 平台打包和套件整合 | 负向 | 当前 | GitHub、GitLab、Snyk 和更宽的 AppSec 平台可借现有合同吃掉预算 | 复盘与打包方案对阵的胜负数据,以及 Socket 到底是增量还是替代 |
这个品类确有监管和威胁侧顺风,但付费供应商仍要顶住低价格下限、买方疲劳和现有平台分发。单靠增长,不足以保证独立经济性好看。
[CM028, CM029, CM030, CM031, CM032, CM033]2.5 图表要点
03竞争格局
3.1 竞争格局概览
Socket 不再只是同另一个 npm 扫描工具竞争。保留来源显示,2026 年买方有四类真实替代。第一类是 Snyk、Mend、Endor Labs、 JFrog Xray 和 FOSSA 等直接专业对手;它们都争夺依赖、软件成分分析或修复预算,只是可达性、合规和平台广度组合不同。 第二类是 GitHub 原生替代品:Dependabot 和 GitHub Advanced Security 已经位于许多团队发现并修复依赖问题的代码库工作流中, 因而成为 Socket 必须打败的默认基线。第三类是更广的 code-to-cloud 或 ASPM 平台,如 Aikido、OX、Apiiro 和 Upwind; 它们把 SCA、SBOM、CI/CD、云、API 或运行时上下文打包进单一合同。第四类是现状:复杂度较低的团队可以用免费依赖告警加内部软件包治理流程达到「足够好」, 从而推迟购买独立专业工具。 这个结构重要,因为 Socket 的差异化真实但窄。Socket 官方页面强调基于行为的恶意软件包检测、安装时拦截,以及以可达性驱动的 CVE 降噪, 而不是全栈 AppSec 或 CNAPP 叙事。独立 2026 年评测覆盖得出类似结论:买方明确关心开发者工作流中的供应链攻击时,Socket 最强, 尤其围绕 JavaScript 和 npm;而更广或更多语言栈的组织有理由考察能整合更多安全栈的供应商。[CP001, CP002, CP005, CP018, CP020, CP027]
| 竞争对手 | 类别 | 与 Socket 的主要重叠 | 定价 / 打包信号 | 最适合买方 | 相比 Socket 的关键短板 |
|---|---|---|---|---|---|
| Socket | 依赖与供应链安全专精厂商 | 基于行为的包风险识别、防火墙和可达性主导分流 | Free '$0'、Team '$25'、Business '$50';Enterprise 按开发者定制 | 希望拦截恶意包并降低 CVE 噪音的开发者优先团队 | 代码、云和运行时覆盖比平台整合型厂商窄 |
| Snyk | 广义 AppSec / SCA 平台 | 依赖,加上代码、容器、IaC、API/web 和 AI 工作流 | Free、Team、Ignite 和 Enterprise 层级,按贡献开发者定价 | 希望整个 SDLC 用一个供应商统一的多语言开发组织 | 安装时恶意软件拦截叙事不如 Socket 专精 |
| Mend | 企业 AppSec 加 Renovate 式自动化 | 可达性驱动 SCA、AI 代码控制和依赖自动更新 | 报价主导、按贡献开发者计费 | 希望一个供应商覆盖代码、依赖和修复自动化的大型 AppSec 项目 | 公开价格透明度较低,围绕包行为的 Socket 式专精品牌也较弱 |
| Endor Labs | 可达性优先 / AI 原生 AppSec | 全栈可达性、有证据支撑的发现和策略定制 | 免费开发者工具;企业平台按演示 / 报价 | 被误报和优先级噪音淹没的企业 | 公开价格透明度较低,安装时拦截叙事也不如 Socket 聚焦 |
| JFrog Xray | 以 artifact 和 registry 为中心的 SCA 老牌厂商 | JFrog Platform 内的 repo、构建、容器、SBOM、许可证和恶意包扫描 | 打包进 Pro X、Enterprise X 和 Enterprise+ 平台层级 | 围绕 Artifactory 和 artifact 治理标准化的 DevOps 与平台团队 | registry 是控制点时更强;repo 原生开发者工作流中不如 Socket |
| FOSSA | 合规与许可证平台 | SBOM、许可证、代码片段和二进制扫描,带安全附加能力 | Free、'$20 per project per month' Business、Enterprise 定制 | 法务、合规和审计负担重的组织 | 对安装前恶意包拦截的强调弱于 Socket |
| GitHub Dependabot + GHAS | GitHub 原生替代 | repo 工作流里的已知漏洞告警、代码安全和密钥保护 | Dependabot 内置;GHAS 附加项为每活跃提交者每月 '$19' 和 '$30' | 以 GitHub 为中心、想要最低摩擦默认项的团队 | 依赖行为专业深度不如 Socket,但更原生、更方便 |
| Apiiro | ASPM 与软件供应链平台 | 风险图、上下文 SCA、安全内建和 XBOM 生成 | 演示主导的平台销售 | 需要深应用上下文和程序化工作流的安全团队 | 间接替代,而非纯依赖安全专精厂商 |
| Aikido | 统一的开发者到运行时安全平台 | 一个平台覆盖 SCA、SAST、IaC、DAST、容器、云、运行时和恶意软件检测 | 公开定价,含免费层和企业附加项 | 希望用一份合同整合多个单点工具的团队 | 覆盖更宽,但不如 Socket 聚焦包行为这一细分场景 |
| OX Security | 代码到云 AppSec 平台 | 一份许可证覆盖 SAST、SCA、SBOM、CI/CD 安全、运行时和渗透测试 | 一个平台、一个价格,按开发者定价 | 正在整合扫描器和交付栈控制的成熟 AppSec 项目 | 在依赖安全专精采购中,品类心智不如 Socket 直接 |
| Chainguard / Upwind | 相邻供应链和运行时替代品 | 可信镜像和库,或以运行时为先的云与 AI 安全,并带 SCA/SBOM | Chainguard 的 10 人 Catalog 团队起价 '$19K';Upwind 演示主导 | 受监管且容器负载重的团队,或向左扩展的云安全买方 | 替代价值来自不同控制点,而非 Socket 以 repo 为中心的工作流 |
公开定价和打包单元格只使用已保留的现行供应商页面;报价主导类别直接标注,不做估算。
[CP002, CP006, CP009, CP011, CP014, CP016]基于证据的序数图:横轴是平台广度和打包能力,纵轴是依赖专用的恶意包识别与分诊深度。
两条轴都是分析师按 1 到 5 给出的序数估计,依据保留的官方和独立来源,并非审计基准。横轴越高,意味着整合范围或工作流卡位更强;纵轴越高,意味着依赖专用信号越深,尤其是恶意包识别或以可达性牵引的分诊。
[CP005, CP006, CP011, CP014, CP018, CP023]3.2 直接专业厂商与既有平台
Snyk、Mend、Endor Labs、JFrog Xray 和 FOSSA 是最接近的直接竞争对手,因为每一家都有可能赢下同一笔开源风险降低预算, 尽管解决方式并不相同。Snyk 销售的是保留集合里最广的开发者优先平台,覆盖 SCA、代码、容器、IaC、API 和 AI 工作流, 并带有可达性感知优先级和自动修复 PR。Mend 同样推更大的平台逻辑,在按贡献开发者计费模型下加入可达性驱动 SCA、 AI 代码控制和 Renovate 式依赖自动化。Endor Labs 对 Socket 的分流叙事压力最大:它营销横跨一方代码、依赖和容器镜像的全栈可达性, 其竞争页面则攻击 Socket 在策略上不够透明、可定制性较弱。 JFrog Xray 和 FOSSA 重要,但原因不同。Xray 最强的场景是控制点在制品流水线而不只是代码库,因为它会在 JFrog Platform 内持续扫描代码库、 构建包、镜像和已存制品。FOSSA 更少关注抓住下一轮 npm 恶意软件活动,更多关注合规运营、SBOM、代码片段扫描和二进制扫描。 这些直接对手合在一起说明,Socket 的竞争并不只是「另一个扫描器」:有的供应商拼开发者工作流广度,有的拼可达性深度, 还有的拼合规或制品治理成熟度。[CP006, CP007, CP008, CP009, CP010, CP011]
| 能力 | Socket | Snyk | Endor Labs | JFrog Xray | GitHub Dependabot / GHAS | Aikido / OX |
|---|---|---|---|---|---|---|
| 安装前恶意包拦截 | 强 | 中 | 中 | 中 | 弱 | 中 |
| 可达性 / 可利用性上下文 | 强 | 强 | 强 | 中 | 弱 | 中 |
| 许可证、SBOM 和合规运营 | 中 | 强 | 中 | 强 | 弱到中 | 强 |
| 代码、容器、IaC 或运行时覆盖宽度 | 中 | 强 | 强 | 强 | 中 | 强 |
| 原生工作流分发 | 中 | 中 | 中 | 弱 | 强 | 中 |
| 策略和治理可扩展性 | 中 | 强 | 强 | 强 | 中 | 强 |
单元格只总结已保留的公开证据。'Strong' 表示供应商在保留来源中明确把该能力放在中心;'weak to moderate' 表示有部分证据,但不是品类领先。
[CP001, CP005, CP007, CP011, CP014, CP019]| 供应商 | 公开标价信号 | 合同模式 | 覆盖范围 | 未知项 / 注意事项 | 战略含义 |
|---|---|---|---|---|---|
| Socket | $0 / $25 / $50 每开发者;Enterprise 定制 | 席位制 SaaS,同一套餐结构内可单独购买产品 | 依赖安全、防火墙、可达性、SBOM、GitHub Actions、AI 模型扫描 | 实际企业折扣未披露 | 透明的专精定价利于小团队起步,但相对原生基线,付费支出是增量预算 |
| Snyk | 公开 Free / Team / Ignite / Enterprise 层级 | 按贡献开发者 | 广义 AppSec 平台,覆盖 SCA、代码、容器、IaC、API/web | Team 和 Ignite 的具体单位经济性随产品组合而变 | 为广义开发者优先 AppSec 提供强公开价格锚点 |
| Mend | 已保留页面没有公开单价 | 按贡献开发者 | AppSec 平台,加 AI 高级功能和 Renovate 式自动化 | 报价主导定价限制横向可比性 | 广义平台买方需要单独尽调实际标价到净价的落差 |
| FOSSA | Free 加 '$20 per project per month' Business | 按项目计费,并向 Enterprise 上售 | 许可证、漏洞、SBOM、代码片段和二进制工作流 | 安全深度取决于附加项和企业配置 | 面向法务和审计买方的透明合规价格锚点 |
| GitHub Dependabot + GHAS | Dependabot 告警内置;Secret Protection 每活跃提交者每月 '$19',Code Security 为 '$30' | GitHub 工作流内附加项 | 依赖监控、代码安全和密钥保护 | Dependabot 范围限于已知漏洞覆盖和受支持生态 | 在以 GitHub 为中心的账号中形成最清晰的默认价格锚点 |
| Chainguard | Catalog 中 10 人团队起价 '$19K' | 按工程组织规模或镜像 / 生态范围授权 | 加固镜像和库,带 CVE 修复 SLA | 替代经济性取决于多少支出从 repo 转向镜像和库 | 受监管、容器负载重买方的强相邻锚点 |
| Aikido | 公开定价,含免费层和企业服务 | SaaS 套餐,可选本地部署和设备 / 运行时扩展 | 代码、云、运行时、SCA、SAST、IaC、DAST 和恶意软件检测 | 高级服务仍适用 Enterprise 定制条款 | 透明的整合主张可抑制专精工具蔓延 |
| OX Security | 公开话术是 'one platform, one price, one license' | 按开发者 | 代码到云平台,覆盖 SAST、SCA、SBOM、CI/CD、运行时和渗透测试 | 已保留来源未披露公开数字标价 | 整合故事比标价透明度更重要 |
| JFrog Xray | 定价嵌入 Pro X、Enterprise X 和 Enterprise+ 订阅 | 平台打包订阅 | repo、构建、容器、artifact、SBOM 和合规扫描 | 已保留定价来源未披露清晰的 Xray 单独单价 | Artifactory 已深度扎根时,平台打包会很有吸引力 |
本表刻意区分公开数字锚点与报价主导或平台打包方案。缺失标价仍是尽调事项,而不是估算。
[CP002, CP008, CP010, CP015, CP017, CP018]3.3 打包替代品与相邻替代品
GitHub 是最重要的打包替代品,因为许多团队已经在它拥有的工作流里审查依赖、PR、告警和修复工单。 GitHub Advanced Security 明确营销内置密钥保护、代码安全和依赖监控,Dependabot 告警则直接在代码库中覆盖已知脆弱依赖。 这种原生位置创造了 Socket 单靠自己难以匹配的分发力量:即使 GitHub 的依赖覆盖没有 Socket 的行为模型那么专业,采购和工作流摩擦也更低, GitHub 公开的按活跃提交者计价还给买方提供了「足够好」防护的清晰基线。 相邻平台带来另一种风险。Apiiro、Aikido、OX Security 和 Upwind 都在营销比 Socket 更广的上下文:基于图的或 code-to-cloud 可见性、 安全内建策略、API 与运行时覆盖,或者用一个平台替换多个扫描器。Chainguard 则从另一个角度竞争, 把控制点移到带有合同化 CVE 修复 SLA 的加固镜像和库上。它们不是一对一的 Socket 克隆,但只要买方偏好整合、 云 / 运行时上下文或制品来源,而不是专业依赖安全工具,它们就是可信替代。因此,Socket 的真实竞争越来越包括想吸收供应链支出的平台, 而不只是看起来像 Socket 的对手工具。[CP018, CP019, CP020, CP021, CP022, CP023]
| 替代类别 | 默认控制点 | 买方为何留在原位 | Socket 反向定位 | 切换 / 多归属动态 |
|---|---|---|---|---|
| GitHub 原生基线 | 仓库、pull request 和安全标签页 | 已嵌入日常开发工作流,并有公开附加项定价 | 专精恶意包检测和更深的依赖风险信号 | GitHub 分发优势高;Socket 可以叠加,但必须证明额外支出合理 |
| 广义 AppSec 平台 | 现有 SAST 或平台安全合同 | 一家厂商可覆盖依赖、代码、容器和策略 | 专精厂商在供应链攻击和安装时拦截上的叙事更清晰 | 多家并用很容易,Socket 可能拿到一个席位,不必替换平台厂商 |
| 可达性优先的专精厂商 | 漏洞积压和减少误报的采购动机 | 买家更想要可利用性上下文和修复工作流,而不是新的扫描器 | Socket 可把行为信号与自身可达性能力配在一起,但必须证明工作流影响相当 | 噪声高的企业环境里,竞争重叠最高 |
| 合规 / 审计权重高的既有厂商 | SBOM、许可证或制品治理流程 | 法务和监管结果可能主导购买标准 | 当供应链恶意软件风险比文档和归因工作流更重要时,Socket 更容易赢 | 切换意愿通常较低,除非合规成为首要用例 |
| 代码到云平台 | 一个控制台覆盖 AppSec、CI/CD、云、API 和运行时可见性 | 整合会降低控制台疲劳和采购负担 | Socket 提供更深的专精依赖能力,不提供全栈覆盖 | 专精厂商最难打的账户,往往明确要削减工具数量 |
| 内部自建 / 维持现状 | Dependabot 式告警加内部包审查政策 | 够用的基线免费,或团队已经有人维护 | Socket 减少基于行为的攻击和安装前拦截盲区 | 切换成本最低,但购买紧迫性也最低 |
本表比较的是控制点和购买惯性,不单独评价产品质量;目的在于看清哪些场景里 Socket 要对抗既有厂商的天然占位,哪些场景可以叠加切入。
[CP019, CP020, CP031, CP035, CP041, CP042]买家匹配矩阵,展示 2026 年 Socket 所在品类最关键的五个决策维度中,各家厂商的强项。
评级只汇总保留证据。强表示保留来源多次明确支持;中表示有邻近能力但不是核心;弱表示该维度并非公开主强项。
[CP021, CP023, CP025, CP026, CP032, CP037]3.4 护城河耐久性、切换成本与风险
Socket 仍有可防守的护城河,但这是专业工具护城河。最强的保留证据支撑两个差异化控制点:代码运行前基于行为拦截恶意软件包, 以及由可达性驱动的降噪,现在还延伸到全应用分析。这些能力对明确担心供应链攻击的团队最重要,而不只是追求安全公告新鲜度。 买方已经感受过受损软件包、拼写仿冒或压倒性 CVE 噪声的痛苦时,Socket 的故事清晰且可信。 风险在于,这条护城河不会自动转化为账号的独占所有权。多家并用完全可能,因为 Socket 可以同 GitHub、Snyk 或更广 AppSec 工具并排存在; 这会降低切换成本,也可能限制定价权。2026 年公开评论还显示,简单漏洞检测正在商品化,因此买方越来越会问: 是否应该整合进 GitHub、Snyk、Mend、Aikido、OX 或 Upwind,而不是再加一个专业工具。关键尽调问题因此不是 Socket 是否存在技术差异化;它确实存在。更难承保的问题是,这种差异化在多大频率上足以压过 GitHub 原生分发、 更广平台合同,以及只购买「足够好」依赖覆盖的现状倾向。[CP004, CP027, CP033, CP034, CP036, CP040]
| Socket 护城河 / 风险 | 为什么重要 | 主要威胁 | 严重程度 | 缓释动作 / 尽调问题 |
|---|---|---|---|---|
| 基于行为的恶意包检测 | 让 Socket 区别于只围绕已知漏洞时效性做文章的工具 | 不把包行为风险放在优先级的买家,可能觉得 GitHub、Snyk 或平台厂商已经够用 | 高 | 量化恶意包担忧有多常成为 Socket 赢单的明确原因 |
| 安装时防火墙 | 把控制点前移到合并后或扫描后补救之前 | 买家可能偏好更少代理,或改用注册表、镜像、工作流控制 | 高 | 验证 Firewall 相比单独核心扫描的加购率和续约影响 |
| 可达性牵引的噪音压降 | 帮 Socket 走出纯恶意软件检测,进入分诊效率 | Endor Labs 和 Snyk 在可利用性上下文与平台广度上的公开叙事更强 | 高 | 专门对标 Endor 和 Snyk,比较 Socket 在大型高噪声环境中的赢率 |
| 透明公开定价 | 团队想从小规模起步时,透明定价降低初始采用摩擦 | GitHub 维持更低摩擦的原生基线,企业标价到实收的机制仍未披露 | 中 | 按客群收集实收价格和扩张路径,检验相对原生替代品的定价权 |
| 专精定位 | 供应链敏感账户里,更尖锐的专精叙事可能跑赢更宽的平台 | Aikido、OX、Apiiro 和 Upwind 可把同一预算吸收到更大的代码到云合同中 | 高 | 衡量整合要求有多常在技术评估前就把 Socket 排除 |
| 适合多家并用 | 叠加部署可帮助快速进入客户 | 叠加容易也会降低切换成本,并可能限制长期钱包份额 | 中 | 索取相对 GitHub 和 Snyk 的模块加购、多产品重叠和替换数据 |
严重程度基于留存公开证据,反映未来 12-24 个月价值捕获承压的可能性,而不是当前流失。
[CP033, CP034, CP036, CP037, CP040, CP041]用保留的公开证据而非管理层指引,压缩呈现 Socket 竞争位置的耐久度。
分数是分析师基于保留证据给出的 0-10 判断,概括耐久度,不是经审计的运营指标。
[CP033, CP034, CP040, CP041, CP043, CP046]3.5 图表要点
04财务情况
4.1 定价模型、牵引力与收入形态
对一家私有安全创业公司而言,Socket 的公开变现界面异常清晰,尽管实际经济性仍不透明。定价页显示按开发者计费的 SaaS 模型: $0 Free 层级、$25 Team 层级、$50 Business 层级和自定义 Enterprise 合同,另有年度预付折扣、创业公司折扣、市场采购和大客户手工开票。 这个结构意味着收入更像经常性订阅,而不是服务驱动变现,但它没有披露实际 ACV、折扣或续约质量。最强牵引力信号来自 Socket 自己的 2026 年 5 月 Series C 材料和首页:超过 27,000 个组织受保护、1.5 million 个代码库、每月 11.6 million 次提交受保护、 每周超过 10,000 次攻击被拦截。这些指标加上 Anthropic、xAI、Replit、Cursor、Figma、Vercel、Gusto、Mercado Libre 和 Cribl 等具名客户,支撑需求和企业适配度。缺失的仍是核心收入账本:没有官方 ARR、GAAP 收入、毛利率或留存披露。[CI001, CI002, CI003, CI005, CI007, CI008]
| 收入流 | 机制 | 公开价格 / 单位 | 当前状态 | 收入质量判断 | 尽调问题 |
|---|---|---|---|---|---|
| 免费 / 开源 | 面向个人、小团队和开源项目的自助入门层 | $0 / 开发者 / 月 | 明确支持免费层和开源使用 | 漏斗好,但不能直接变现 | 按仓库类型和组织规模拆分的免费转付费转化 |
| 团队订阅 | 面向成长团队的付费开发者席位计划 | $25 / 开发者 / 月 | 增加自动化和预计算可达性 | 标价清晰的经常性席位收入,但实收折扣未知 | 团队平均 ACV、席位利用率和年度预付占比 |
| Business 订阅 | 更高端的自助 / 商业计划 | $50 / 开发者 / 月 | 无需强制销售沟通即可获得无限扫描和 API 配额 | 在完整企业合同前承接更大账户 | Business 计划客户数和升级到 Enterprise 的比例 |
| 企业合同 | 带高级可达性和支持的定制企业套餐 | 定制 | 手动开票、Marketplace 购买、指定支持、SCIM、审计日志 | 可能是 ACV 最高的收入流,但实收价格未披露 | 企业 ACV 中位数、合同期限和折扣政策 |
| 多产品加购 | 同一计划体系内可单独购买的附加产品 | 随产品而异 | 定价页列出威胁情报、认证补丁、防火墙、密钥、容器、扩展扫描 | 如果加购率真实,可能抬高 NRR,但加购数据不公开 | 按产品拆分的加购率和增量毛利率 |
| Marketplace / 年度采购 | 不只按月刷卡,也支持年度预付和企业采购路径 | 年度最高节省 20%;GCP Marketplace / ACH-Wire | 支持企业采购和财务流程 | 有助于降低采购摩擦,但不能证明实收回款质量 | 年度条款、银行卡、发票和 Marketplace 各自贡献的开票额占比 |
本表描述公开标价和收入机制,不涉及实收收入确认或 ACV。
[CI008, CI009, CI011, CI012, CI013, CI019]| 计划 / 杠杆 | 公开标价 | 计费单位 | 包含容量 / 信号 | 仍然未知 | 来源 |
|---|---|---|---|---|---|
| Free | $0 | 每位开发者 / 月 | 开发者和仓库不限、1,000 次扫描、恶意软件和许可证检查 | 免费使用转化为付费扩张的频率 | Socket 定价 |
| Team | $25 | 每位开发者 / 月 | 5,000 次扫描、10 名成员、可达性、Slack 告警 | 实收折扣和每个付费团队的平均席位数 | Socket 定价 |
| Business | $50 | 每位开发者 / 月 | 无限扫描、无限成员、合规和 API 功能 | 卡付与年度合同之间的混合实收价格和结构 | Socket 定价 |
| Enterprise | 定制 | 合同 | 全应用函数级可达性、SCIM、审计日志、私有 Slack、指定客户经理 | ACV 区间、期限、最低额和爬坡结构 | Socket 定价 |
| 年度预付 | 最高节省 20% | 年度计费 | 明确的年度 vs 月度取舍 | 采用年度条款的客户占比 | Socket 定价 FAQ |
| 采购灵活性 | 定制 / 企业 | 发票、ACH/Wire、GCP Marketplace | 支持手动开票和 Marketplace 购买 | 回款时点、Marketplace 占比和渠道费用 | Socket 定价 FAQ |
公开标价可见;企业实收价格、折扣和回款质量未披露。
[CI008, CI009, CI010, CI011, CI012, CI013]Socket 变现靠免费 / 开源漏斗、按开发者计费的订阅套餐,以及不透明的企业扩张,而不是一次性安全服务。
[CI008, CI009, CI011, CI019, CI020, CI024]4.2 单位经济性代理与成本结构线索
公开单位经济性证据大多是代理指标,但这些代理指向软件式经济性,而不是重资产模式。Socket 销售云交付分析、席位和高级工作流功能, 而不是硬件、库存或项目融资重的资产。定价页按开发者包装和企业支持功能显示,毛利率更多取决于算力、数据处理、支持和商业化效率, 而不是实体交付成本。同一页面还暴露分层增购路径,例如合规集成、审计日志、SBOM 工作流、可达性、私有 Slack 频道和指定客户经理; 这些都意味着大客户会带来一定增量服务交付负担。Coana 在财务上重要,因为其可达性技术被定位为降低误报和修复时间; 如果这些主张成立,收购可以提高产品价值和留存,同时不改变 Socket 的基本软件成本结构。公司仍未发布 CAC、回本周期、毛利率、支持比率或 NRR, 所以公开分析只能停在方向性推断,而不能形成可防守的单位经济性模型。[CI013, CI019, CI020, CI024, CI025, CI026]
| 指标 | 公开数值 / 状态 | 置信度 | 为什么重要 | 尽调问题 |
|---|---|---|---|---|
| 可计费用户定义 | Developer = 过去 90 天内向已扫描仓库提交代码的人 | 中 | 定义变现分母和席位弹性 | 按队列拆分的付费开发者,以及闲置席位流失 |
| 受保护组织 | >27,000 家组织 | 高 | 公开资料中最好的需求广度信号,可验证漏斗顶部和企业相关性 | 付费 vs 免费组织,以及客群结构 |
| 受保护仓库 | 1.5M 个仓库 | 高 | 显示被监控足迹的规模和基础设施需求 | 每个付费客户的平均仓库数 |
| 每月保护提交 | 每月 11.6M+ 次提交 | 高 | 使用强度和平台依赖的活动代理 | 按计划拆分的提交量与收入相关性 |
| 每周拦截攻击 | >10,000 次供应链攻击 / 周 | 高 | 可证明产品活动,但不直接等于变现 | 付费 vs 免费环境贡献的拦截攻击占比 |
| 公开团队规模信号 | 官方博客称团队 >100 人;ZoomInfo 称 51-200 名员工 | 中 | 人员成本可能是软件安全公司最大的运营费用项 | 按职能拆分的员工数和招聘计划 |
| 第三方收入线索 | ZoomInfo 估算收入约 $18.1M | 低 | 只能为估值提供粗略背景;不是可审计的公司披露 | 按季度收入和 ARR 桥接 |
| 聚合转载的增长说法 | 部分收购报道声称收入同比增长 ~300% | 低 | 可能意味着扩张极快,但来源质量弱 | 显示实际收入增长的董事会材料 |
| 毛利率 / COGS | 未公开披露 | 低 | 验证软件型经济性和支持负担所需 | 按产品拆分的毛利率,以及托管 / 支持成本明细 |
| ARR / GAAP 收入 | 未官方披露 | 低 | 估值测算的核心输入 | 历史 ARR、GAAP 收入、开票额和递延收入 |
| 烧钱速度 / 现金跑道 | 未公开披露 | 低 | 测试融资依赖和下行情景韧性所需 | 月度烧钱、现金余额和悲观情景现金跑道 |
公开单位经济证据大多来自代理指标。等同 null 的行表示,截至 2026-05-24,留存公开来源未找到该数据。
[CI005, CI006, CI010, CI021, CI022, CI023]公开的单位经济叙事依赖活跃度和工作流代理指标,而不是披露的获客成本(CAC)、回本周期或利润率。
[CI021, CI022, CI023, CI024, CI033, CI038]4.3 资本充足性、融资历史与估值含义
融资形成比资本充足性清楚得多。2026 年 5 月官方披露、法律报道和独立报道一致显示,Socket 以 $1 billion 估值完成 $60 million Series C,由 Thrive Capital 领投,Andreessen Horowitz、Abstract Ventures 和 Capital One Ventures 参投。 Socket 自己的博客称,该轮把累计融资带到 $125 million。更早的公开报道显示,2024 年 10 月由 Abstract Ventures 领投的 $40 million Series B 已把累计融资推至 $65 million;分析师数据库则指向 2022 年 5 月的第一轮融资。读下来, Socket 在不到两年里筹集了可观的股权缓冲,股权结构表里也有可识别的风投和天使。但外部仍无法验证现金模型。 公开资料没有现金余额、烧钱速度、现金跑道桥或债务安排。因此,估值看起来更像是押注客户质量、AI 驱动需求和产品广度的信心交易, 而不是一套可公开审计的利润率或现金流故事。Coana 增加了资本用途的战略信号,但即便该交易,对价也未披露。[CI001, CI002, CI003, CI014, CI015, CI016]
| 项目 | 公开数值 / 状态 | 证据基础 | 投资判断含义 | 尽调问题 |
|---|---|---|---|---|
| 最新新股融资 | $60M Series C 轮,2026-05-20 | Socket 官方博客 / 新闻稿,加上 Cooley 和媒体 | 对继续招聘和产品扩张而言,新股资金可观 | 股权结构表、股数和清算优先权 |
| 最新估值 | $1B 投后估值 | 官方 Series C 材料加独立报道 | 对持续增长和最终利润率质量设下高门槛 | 内部运营计划 vs 估值假设 |
| 已披露累计融资 | $125M | 官方 Series C 博客,加上 Cooley 和 Tracxn | 为软件公司提供可观股权缓冲 | 各轮融资款用途和当前现金余额 |
| 上一轮主要融资 | $40M Series B,October 2024 | TechCrunch、Cooley、GlobeNewswire 聚合转载 | 说明 2026 年迈入独角兽估值前已有强后续资金支持 | Series B 到 Series C 的董事会材料 |
| 融资节奏 | ~19 个月,Series B 到 Series C | 来自 October 2024 和 May 2026 两个公开日期 | 指向较快融资推进,并带来更大估值跃升 | 2024-2026 期间月度 KPI |
| 公开员工数信号 | 官方博客称团队 >100;外部估计范围 51-200 | 官方 Series C 博客和 ZoomInfo | 显示持续的运营费用负担,但不是精确烧钱额 | 按职能拆分的员工数和全负荷现金薪酬 |
| 公开资金用途叙事 | 扩展平台、推动企业采用,并保护 AI 时代供应链 | 官方 Series C 材料和媒体复盘 | 支持增长投资逻辑,但不是现金充足证明 | 招聘、R&D、销售和 M&A 整合预算 |
| 手头现金 | 未披露 | 留存公开来源未发布现金余额 | 外部无法建模现金跑道 | 月度现金余额和最低运营现金政策 |
| 烧钱速度 / 现金跑道 | 未披露 | 留存公开来源未发布净烧钱或现金跑道月数 | 无法测试下行情景中的融资依赖 | 月度烧钱桥接和悲观情景现金跑道 |
| 债务 / 老股交易 | 未找到公开债务、风险债务或老股交易条款 | 留存公开来源聚焦股权融资 | 没有证据不等于证据显示没有 | 债务明细表、贷款协议和老股交易 |
| Coana 对价 | 未披露;TFN 猜测 $50M-$100M | 官方和独立收购报道 | M&A 现金使用方向上有参考价值,但仍不透明 | 收购协议、现金 / 股票结构和留任方案 |
公开融资历史比公开流动性更清楚。等同 null 的行表示,截至 2026-05-24,留存公开来源未找到该信息。
[CI001, CI002, CI003, CI014, CI015, CI016]公开测算锚点覆盖估值、累计融资、员工数,甚至 Coana 估计价格区间;但可审计收入或现金没有公开锚点。
员工数口径把团队超过 100 人的官方表述,与 ZoomInfo 的 51-200 人区间合并使用。Coana 对价只是 TFN 估计,公司未确认。
[CI001, CI003, CI006, CI026, CI032, CI034]公开证据指向软件式资本强度,但可见资金用途包括招聘、企业支持、计算和 M&A 整合;现金余额仍未公开。
[CI025, CI026, CI030, CI032, CI040, CI043]4.4 反向视角与披露阻塞点
Socket 的反向财务视角不是可见困境,而是披露质量。公开来源能支撑融资规模、投资人名单、标价、客户标识和增长叙事, 却无法承保收入质量或下行韧性。第三方市场数据网站提供部分替代,但会引入噪声:ZoomInfo 模型估计收入约 $18.1 million、 员工 51-200 人;Tracxn 显示一家已融 $125 million 的 Series C 公司,但隐藏关键数字;Scamadviser 一边报告信任分为零, 一边又说网站可能合法。即便 Socket 自己的界面也并非完全一致,About 页面称公司创立于 2021 年,多份融资材料称 2020 年。 收购经济性同样不完整:官方和独立报道都同意 Coana 于 2025 年 4 月被收购,但收购价格和整合成本未披露; 少数已发布的收入增长或收购价格估算看起来是聚合或投机性质。本章结论因此是建设性但不完整:经常性软件收入看似可信, 资本强度看起来适中,投资人支持强;但严肃承保仍需要只由管理层提供的数据。[CI018, CI027, CI028, CI029, CI030, CI032]
| 缺失指标 / 文件 | 公开状态 | 为什么重要 | 当前代理指标 | 精确尽调路径 |
|---|---|---|---|---|
| ARR 和 GAAP 收入 | 未官方披露 | 对标估值和增长质量所需 | 只有 ZoomInfo 估算和聚合转载的增长说法 | 管理层收入历史、ARR 桥接和递延收入明细表 |
| 毛利率和 COGS | 未公开披露 | 验证软件型经济性和支持负担所需 | 只有定价架构和产品交付推断 | 按产品拆分的毛利率,以及托管 / 支持成本明细 |
| 现金、烧钱速度和现金跑道 | 未公开披露 | 测试融资依赖和下行情景韧性所需 | 只有融资规模历史 | 月度现金桥接、烧钱预测和悲观情景现金跑道 |
| 企业实收价格 / 折扣 | 未公开披露 | 标价无法揭示 ACV、期限或续约质量 | 公开标价加企业功能层级 | Top 20 合同样本,含 ACV、期限、折扣和续约状态 |
| NRR、流失和客户集中度 | 未找到公开证据 | 判断经常性收入耐久性的关键输入 | 具名客户 logo 和组织数量主张 | 净留存、总流失和前 10 大客户集中度 |
| 收购对价和整合成本 | 价格未披露;只有猜测区间 | 需要据此判断现金使用、稀释和协同回本 | 仅有 TFN 估计的 $50M-$100M | 购买协议、现金 / 股票组合、earn-out 条款和整合预算 |
| 债务 / 二级交易义务 | 留存公开来源未找到 | 可能显著改变资本结构风险和流动性需求 | 留存来源均未公开说明 | 债务期限表、认股权证覆盖和任何老股转让计划 |
这张表刻意记录证据缺口,而不是猜测。每行都列出补上承销缺口所需的具体尽调请求。
[CI020, CI027, CI029, CI030, CI032, CI034]4.5 图表要点
05产品与技术
5.1 产品定义与开发者工作流
Socket 最好理解为开发者工作流安全平台,而不是单一静态扫描器。公开入口是一个 GitHub 应用,会监控 PR 中的依赖变更、 在合并前评论,并围绕新软件包产出项目健康度式结果。但这只是一个表面。同一产品族还包括 CLI,用于更自定义或非 GitHub 工作流; VS Code 扩展,把清单文件扫描带进编辑器;以及 Socket Firewall,它位于软件包管理器和注册表之间,把执行点移到安装时。 结果是,产品能在多个时刻看到依赖风险:代码编辑时、PR 提出依赖时,以及软件包真正下载到开发者电脑或 CI 运行器时。 这种多表面设计重要,因为 Socket 销售的是嵌入正常开发者工具链的防护,而不是要求安全团队运行一份孤立的事后报告。[CE001, CE002, CE003, CE004, CE005, CE006]
| 用户任务 | 当前工作流 | Socket 触点 | 可衡量收益 | 限制 |
|---|---|---|---|---|
| 合并前审查新依赖 | 提交包含 manifest 或 lockfile 变更的 PR | GitHub app PR 评论和项目健康报告 | 风险在合并前暴露,而不是部署后才出现 | 受当前集成最擅长解析的生态和包格式限制 |
| 编码时检查 manifest | 在 VS Code 中编辑包文件 | VS Code 扩展 | 开发者不离开编辑器就能看到包级风险 | 不是每条分析路径都完全离线或本地化 |
| 安装时阻止恶意下载 | 运行 npm、pip、cargo 或类似安装命令 | Socket Firewall | 在笔记本或 CI runner 执行前阻断高风险包 | 包管理器路径必须接入安装拦截 |
| 降低 CVE 分诊噪音 | 审查某项服务的漏洞积压 | 可达性层级 | 剔除不可达发现,优先处理可利用路径 | 更高精度层级需要更多配置和计算 |
| 自动化组织专属检查 | 搭建定制安全或报告工作流 | REST API 和 SDKs | 让团队把 Socket 嵌入内部工具和策略流程 | 需要工程投入,小团队未必愿意承担 |
| 跨生态跟踪供应链攻击行动 | 跟进新发布的攻击研究和检测更新 | Socket 研究信息流和包情报 | 在传统披露周期跟上前,提高威胁感知 | 公开文章不披露完整内部检测流水线 |
用例表概括了外部文档记录的主要开发者和 AppSec 工作流,而不是每一种企业部署变体。
[CE003, CE004, CE006, CE007, CE008, CE009]从依赖变更到策略决策的典型开发者到 AppSec 流程。
该流程把 GitHub、编辑器、Firewall 和可达性界面压缩成一条代表性运营路径。
[CE003, CE006, CE008, CE021, CE044]5.2 产品表面与生态地图
产品地图很宽,但证据显示,它是分层变宽,而不是一个单体 SKU。公开材料描述了 GitHub 审查、编辑器时提示、安装时策略执行、 可达性分析、API 与 SDK 访问,以及研究驱动的软件包情报层。Socket 还营销对主流开源注册表的多生态支持; Firewall 和发布帖子显示,公司扩展到了 Maven、Ruby、NuGet、Packagist、OpenVSX,以及 PHP 或 Composer 表面。 同时,公开证据并没有显示每个生态都有完全均匀的深度。GitHub 页面仍最明确地突出 JavaScript、Python 和 Go;FAQ 声称更广语言支持; 独立评测仍把最深适配描述为 JavaScript 中心工作流最强。这意味着,产品组合应被理解为宽且仍在扩张, 但不是一张完全对称的功能矩阵,无法清楚证明每个生态都有相同的行为分析、可达性深度和工作流覆盖。[CE010, CE011, CE012, CE013, CE014, CE015]
| 模块 / 资产 | 主要用户 | 当前角色 | 当前状态 | 差异化 | 主要限制 |
|---|---|---|---|---|---|
| Socket for GitHub 集成 | 应用和平台工程师 | PR 阶段依赖审查和健康报告 | 成熟 / 核心入口 | 安全评论直接进入合并工作流 | 生态深度最明确的仍是 JS、Python 和 Go |
| CLI + API + SDKs | 平台和 AppSec 团队 | 自定义扫描、自动化和集成 | 成熟 / 仓库活跃 | 定制工作流里,比 GitHub app 更可控 | 比点击式安装更吃工程投入 |
| VS Code 扩展 | 编辑器内的开发者 | 编码时扫描 manifest 并审查告警 | 活跃 / 已上线 | 把依赖审查带进编辑器,减少上下文切换 | 部分分析仍依赖 API 连接 |
| Socket Firewall | 开发者效率、平台和安全团队 | 安装时拦截、阻断、遥测和策略执行 | 扩张中 / 高优先级产品线 | 把防护从告警前移到安装时预防 | 企业功能深度比公开社区遥测细节更清楚 |
| 可达性 | 负责漏洞分诊的 AppSec 和平台团队 | 在依赖级、预计算和全应用层过滤不可达 CVE | 扩张中 / 2025-2026 年重大扩展 | 精准分诊补强恶意包检测 | 全应用模式带来不小的配置和运行成本 |
| 研究引擎和包情报 | 安全团队和产品检测 | 跨生态威胁发现、红旗分类和信号更新 | 核心支撑层 | 研究到产品的闭环强化新型威胁覆盖 | 公开研究数量本身不能证明各生态产品深度相同 |
矩阵按外部可见产品面归类,而不是穷举每个 SKU 或企业计划组合。
[CE002, CE005, CE006, CE008, CE010, CE011]从开发者入口到分析和执行层的代表性技术栈。
这是基于官方文档和发布说明拼出的公开运营架构,不是内部服务图。
[CE002, CE013, CE019, CE020, CE044]5.3 架构与可达性运营模型
Socket 的运营架构从依赖摄入开始,而不是运行时遥测。它消费清单文件、锁文件和安装请求;分类行为和元数据; 纳入维护者行为启发式;再把情报投射回开发者面向的表面。可达性现在是这套栈之上的主要精度层。公开产品页面区分三层: 依赖可达性、预计算可达性和完整应用可达性。低摩擦层级可以跨现有集成运行;完整应用可达性更重,需要 CLI 或 GitHub Action 设置以及特定语言运行时前置条件。 Coana 收购是这里的关键技术动作:Socket 明确使用 Coana 的静态分析和控制流分析来判断某个 CVE 是否真正可达; 官方材料强调,预计算可达性可以在不立刻要求上传源代码的情况下带来大幅降噪。这是相对于只停留在漏洞匹配列表工具的实质差异化。[CE019, CE020, CE021, CE022, CE023, CE024]
| 层 / 组件 | 作用 | 依赖 | 关键风险 | 证据 |
|---|---|---|---|---|
| Manifest 和安装入口 | 收集 manifest、lockfile 和安装请求,作为原始包图输入 | 包管理器和代码仓库集成点 | 入口漏掉某个依赖面后,下游检测补不回来 | GitHub、Firewall 和文档页面 |
| 行为分析引擎 | 检查网络、文件系统、shell、环境、安装脚本、遥测和混淆信号 | 语言解析器和包源码可见性 | 行为分析仍可能带来调优工作或误报 | FAQ 加独立评测 |
| 元数据和维护者启发式 | 用元数据变化、维护者行为和发布异常补充风险信号 | 注册表元数据质量和历史包记录 | 信号质量随生态历史和维护者可见性而变 | FAQ |
| 可达性精度层 | 过滤漏洞告警,聚焦可达且可利用路径 | 静态和控制流分析,加上仓库或 CI 执行上下文 | 精度越高,配置和计算成本越高 | 可达性功能页和文档 |
| 源自 Coana 的全应用分析 | 为 CVE 分诊加入函数级和预计算可达性 | Coana 引擎集成和特定语言运行时 | 各生态功能是否对齐,公开材料还未说清 | 收购文章和可达性文档 |
| 报告和执行触点 | 将结果送回 PR 评论、编辑器 UI、组织仪表盘、API 和安装时阻断 | GitHub 权限、编辑器扩展设置、包管理器 hook 和 API 可用性 | 工作流价值取决于团队能否把这些触点一起部署好 | 入门、GitHub、VS Code 和 Firewall 页面 |
架构表反映公开材料可见的运营模型,不是内部服务图。
[CE013, CE014, CE019, CE020, CE022, CE023]外部可见依赖决定 Socket 能否交付预防和精准分诊。
这张图强调公开依赖和约束,不覆盖每一个内部服务依赖。
[CE009, CE023, CE024, CE026, CE027, CE042]5.4 发布节奏、路线图与开发者信号
Socket 在 2025-2026 年的公开发布节奏很快,而且异常可见。产品页面显示相邻发布,包括 OpenVSX 扫描、Ruby 可达性 beta、 Immutable Scans、PHP 和 Composer 支持、Jira、Data Exports;研究信息流同时显示跨 npm、Go、NuGet、PyPI、RubyGems、 Packagist 和扩展生态的大量威胁情报帖子。这个组合说明,公司试图把一方研究转化为发货中的产品表面,而不是把研究当作脱离产品的营销活动。 公开开发者信号也指向同一方向。GitHub 组织显示几十个公开代码库,CLI、VS Code 扩展、JavaScript SDK 和 Python SDK 在接近本报告日期时都有活动。这个信号不是巨大的开源规模,但真实且当前。与此同时,Series C 材料把路线图框在更广的安装时防护、 精准可达性以及扩展和 AI 工具等相邻表面上。[CE028, CE029, CE030, CE031, CE032, CE033]
| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 产品含义 | 来源 |
|---|---|---|---|---|
| 2025(公告) | Coana 收购和可达性集成 | 已上线,集成路径推进中 | 加入静态和控制流可达性,降低漏洞噪音 | Socket 和 Coana 文章 |
| 2025-11-20 | OpenVSX 扩展扫描 | 已上线 / 已公告 | 防护从软件包扩展到开发者工具扩展 | 产品新闻流 |
| 2025-11-17 | Ruby 可达性测试版 | 测试版 | 显示可达性正在走出最初的核心语言叙事 | 产品新闻流 |
| 2025-12-17 | Docker Hardened Images 中的 Firewall | 已上线 / 已捆绑 | 把安装时防护带入加固构建环境 | 产品新闻流 |
| 2026-01-23 | Immutable Scans | 已上线 | 提升审查工作流的结果一致性和可复现性 | 产品新闻流 |
| 2026-02-17 | PHP 和 Composer 支持 | 已上线 | 把包安全叙事扩展到 Packagist 和 PHP 团队 | 产品新闻流 |
| 2026-04-20 | Socket for Jira 集成 | 已上线 | 把告警审查接入工单和修复工作流 | 产品新闻流 |
| 2026-04-23 | Data Exports | 已上线 | 让团队把告警数据迁入自有存储和分析栈 | 产品新闻流 |
发布表使用公开产品新闻流和 Coana 公告展示对外产品节奏;不尝试还原内部路线图。
[CE025, CE028, CE029, CE036, CE037]定性展示公开证据最强的位置,以及能力一致性或保障细节更薄的地方。
单元格是基于已获取的官方和独立来源给出的定性判断,不是内部 KPI。
[CE010, CE011, CE012, CE021, CE041, CE042]5.5 差异化、信任与技术风险
Socket 最清楚的差异化在于,它要抓恶意依赖行为,并在正常开发者工作流里优先处理可达漏洞。相比经典 CVE 扫描, 这个主张更尖锐;GitHub 审查、编辑器反馈、安装时拦截和可达性分流组合在一起,也给了它连贯的平台故事。信任姿态方向上正面, 但公开证据仍只部分透明。官方页面明确表示,专有源代码应留在开发者机器或 CI 环境里,主要同 Socket 共享的是清单文件或依赖列表。 这有助于采购对话,但不等于一套公开保证材料。因此,主要技术风险集中在执行和透明度,而不是产品表面积: 跨语言一致性没有完整文件化,基于行为的分析在动态代码库里仍可能需要调优,需要采购级保证证据的买方也需要看到超过当前公开材料的内容。[CE038, CE039, CE040, CE041, CE042, CE043]
| 控制 / 保证信号 | 状态 | 范围 | 公开证据 | 缺口 |
|---|---|---|---|---|
| 源码边界 | 公开表述 | 客户源码留在开发者机器或 CI 环境 | 定价页和 FAQ | 没有公开架构或审计材料证明该主张在运营中成立 |
| PII 处理边界 | 公开表述 | 服务称不处理 PII 或客户私有信息 | FAQ | 没有公开数据处理附录或信任材料包 |
| PR 阶段治理 | 公开演示 | GitHub app 可在合并前审查依赖变更 | 功能页和应用市场 | 公开文档未展示标准检查之外的详细策略审批工作流 |
| 安装时策略执行 | 公开演示 | Firewall 会阻断、警告并遥测安装尝试 | Firewall 功能页 | 没有按生态或仓库类型拆分的公开误拦截基准 |
| 人工验证加 AI 辅助分析 | 公开声称 | 平台在现代威胁分诊中结合自动化分析和人工验证 | Series C 新闻稿 | 运营人员配置模型和审查 SLA 未公开 |
| 采购级保证材料 | 本证据集中未公开 | 应覆盖认证、渗透测试、审计报告和控制映射 | 抓取的公开页面均缺失 | 买方仍需私下尽调材料来验证企业级信任姿态 |
这张表聚焦公开可见的保证信号。缺失的采购材料按尽调缺口记录,不当作隐藏利好。
[CE038, CE039, CE040, CE041, CE042, CE043]5.6 图表要点
06客户情况
6.1 客户基础与买方画像
Socket 的客户证据指向一种以开发者为中心、但由安全预算牵引的购买动作。最强公开背书集中在 AI 原生、云、身份、合规和安全意识较强的软件组织, 而不是广泛线下企业 [CU002, CU004]。2026 年 5 月官方材料称,Socket 保护超过 27,000 个组织、1.5 million 个代码库和每月 11.6 million 次提交,但这些是平台足迹指标,不是已披露的付费客户数 [CU001, CU024, CU026]。Socket 融资公告和新闻稿突出展示的具名客户标识 包括 Anthropic、xAI、Replit、Cursor、Vercel、Figma、Gusto、Mercado Libre 和 Cribl,另有金融服务和全球媒体中的未具名 Fortune 100 公司 [CU002]。 从案例研究看,预算所有者通常是 CISO、安全负责人、安全工程负责人或平台安全负责人;日常用户则是开发者和平台工程师, 他们直接在现有工作流里收到 PR 评论、GitHub 检查或 API 驱动的批准结果 [CU005, CU006]。部署动作明显轻量: Replit、JumpCloud、SHI、Render 和 GitHub Marketplace 材料都把 GitHub App 或 GitHub 检查推出描述为初始切口, 变更管理负担很小 [CU007, CU041]。这种低摩擦动作似乎尤其匹配高速推进的工程团队:它们在意供应链风险,却负担不起大型手工审查队列。 因此,公开分层与其说按公司规模,不如说按代码强度和治理压力。AI 实验室需要更快依赖批准,云平台想要更低噪声的依赖卫生, 合规供应商需要审计证据,开源或加密项目则需要能处理大型依赖树和贡献者规模的审查工具 [CU008, CU012, CU016, CU021, CU042]。
| 分群 | 买方 / 用户 / 付款方 | 代表性证据 | 主要用例 | 战略价值 | 关键缺口 |
|---|---|---|---|---|---|
| AI 研究实验室和智能体构建者 | 买方:CISO / 安全工程;用户:研究员和基础设施工程师;付款方:中央安全预算 | Anthropic、xAI 与 Cursor | 更快批准新依赖,同时不丢失零日供应链可见性 | 战略价值高,因为 AI 编码和研究速度会放大第三方代码风险 | xAI 和 Cursor 在公开证据中只有 logo;没有合同或部署细节 |
| 开发者工具和云平台 | 买方:安全负责人 / 安全主管;用户:平台和应用开发者;付款方:平台工程或安全 | Replit, Vercel, Render | 内嵌 PR 阶段防护、依赖卫生、monorepo 控制、开发者友好上线 | 契合度强,因为 GitHub 原生部署能把快速发布周期的阻力降到最低 | 合同规模或续约经济性的公开证据很少 |
| 合规、身份和安全平台 | 买方:CISO / 安全工程总监;用户:AppSec、DevSecOps、开发者;付款方:安全 / 合规 | JumpCloud, Drata, JupiterOne | 可达性、许可证策略、SBOM、CI/CD 执行、审计证据 | 重要性在于 Socket 从扫描扩展到治理和客户保证工作流 | 多数证据是筛选过的案例研究内容,而非独立采购证据 |
| 受监管软件和医疗健康平台 | 买方:产品安全 / 安全运营;用户:开发者和合规利益相关方;付款方:安全或平台组织 | Cedar, Doctolib, Gusto | 低噪音依赖安全,加合规支持和审计友好证据 | 在重视可审计性和患者 / 金融数据控制的场景中显示契合 | Gusto 只有 logo;公开证据远薄于 Cedar 或 Doctolib |
| 开源、加密和 web3 生态 | 买方:平台 / 安全负责人;用户:OSS 维护者和资深审查者;付款方:平台 / 安全 | Chia, MetaMask | 审核社区贡献依赖和大型依赖树,不被人工审查卡住 | 有助证明 Socket 能在高度公开、贡献者众多的代码库中运行 | 公开记录仍集中在 JavaScript 占比较高的环境 |
| 大型互联网平台和企业技术集团 | 买方:安全、平台或产品领导层;用户:开发者和审查团队;付款方:中央工程或安全 | Mercado Libre、SHI、Fortune 100 金融 / 媒体 logo | 集中式依赖筛查、最小权限上线和早期威胁阻断 | 暗示 Socket 能越过创业公司规模,卖进大型工程组织 | Mercado Libre 和未具名 Fortune 100 缺少公开部署深度或量化结果 |
行按购买语境和工作流分组,而不是按披露 ARR,因为 Socket 不公布分群收入结构。只有 logo 的账户与有案例深度的账户分开处理。
[CU002, CU004, CU005, CU006, CU016, CU023]| 指标 | 数值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| 受保护组织 | >27,000 | 2026-05 | Socket Series C 文章 / 佐证新闻 | 高 | 漏斗顶部触达大,认知增长快 | 未披露其中多少是付费客户,多少是免费、OSS 或事件驱动用户 |
| 受保护仓库 | 1.5 million | 2026-05 | Socket Series C 文章 | 高 | 暗示安装基础深入多个工程工作流 | 没有仓库到客户或付费账户的映射 |
| 每月受保护提交 | 11.6 million | 2026-05 | Socket Series C 文章 | 高 | 暗示是重复工作流使用,而非一次性扫描 | 未按客户、分群或受保护代码总交易额拆分 |
| Series B 后增长 | 7,500 到 >27,000 家组织 | 2024-10 至 2026-05 | Socket Series C 文章 / TFN | 高 | 支撑其在 AI 驱动开发窗口中的强劲广度扩张 | 公司未披露起点或当前付费客户转化率 |
| Axios 事件带来的入驻峰值 | 24 小时内 >2,000 家组织 | 2026-05 | Socket 新闻稿 / Series C 文章 / Techstartups | 高 | 显示事件驱动获客和强品类相关性 | 未披露这些账户中有多少转化为长期客户 |
| 2026 年 5 月材料中的具名客户 logo | 9 个具名 logo,加未具名 Fortune 100 金融 / 媒体账户 | 2026-05 | Socket 新闻稿 / Series C 文章 | 中 | 品牌信号不错,尤其在 AI 原生账户中 | logo 数量不等于生产深度、留存或收入集中度 |
这张表把平台足迹指标和具名 logo 可见度放在一起,因为 Socket 不公布标准客户漏斗披露。计数不应被解读为付费企业客户数。
[CU001, CU002, CU024, CU025, CU026]公开案例显示,Socket 的可复制路径是从 GitHub 原生试点切入,再扩展到治理和合规。
阶段来自公开案例研究、应用市场和文档材料的综合梳理。公开资料没有披露时长和内部采购步骤;本图展示顺序,不表示耗时。
[CU005, CU006, CU007, CU008, CU012, CU016]6.2 具名验证与部署深度
Socket 的公开验证在不同客户之间强弱差异明显。Anthropic、Replit、Vercel、Cedar、Chia、JumpCloud、Render、 Doctolib、Drata、MetaMask、SHI 和 JupiterOne 都有具名案例研究或详细证言,描述部署表面、买方角色或运营结果 [CU008, CU010, CU012, CU013, CU015, CU016, CU018, CU019, CU020, CU021, CU022]。Anthropic 是量化最好的一条验证: 公司称 Socket 被嵌入内部依赖批准流水线,把手工依赖审查减少 95%,每周为安全工程师节省超过五小时 [CU008, CU009]。 Cedar 和 Chia 提供次强量化证据,分别描述上线后告警负担或未关闭安全告警减少 70% [CU014, CU015]。 Replit 和 Vercel 解释了为什么 Socket 能打动 AI 原生和开发者工具客户。Replit 描述了 GitHub 检查推出、更少误报, 以及围绕传递依赖和合规工作流的更高信心 [CU010, CU011]。Vercel 强调 pnpm 和 monorepo 适配、分阶段推出、 依赖蔓延减少,以及开发者认知负担降低 [CU012]。JumpCloud、JupiterOne、Doctolib、SHI 和 Render 把故事进一步推向治理和运营: SBOM 与许可证支持、CI/CD 执行、审计准备、最小访问部署,以及持久低摩擦的 PR 使用 [CU016, CU017, CU018, CU019, CU022, CU040]。 限制在于,Socket 最亮眼的客户标识并非同等充分。xAI、Cursor、Figma、Gusto 和 Mercado Libre 出现在官方客户名单中, 但已审阅的公开语料没有披露这些账号的部署架构、合同范围或结果 [CU023]。这意味着 Socket 可以合理声称拥有标志性客户标识; 但在案例研究集合之外,公开记录仍更接近客户标识验证,而不是深度生产验证。
| 客户 | 分群 | 部署 / 用例 | 生产环境 / 试点 | 成效 | 限制 |
|---|---|---|---|---|---|
| Anthropic | AI 研究 / 基础设施 | 将 Socket API 嵌入内部依赖审批流程,设分数阈值,并保留人工审核兜底 | 生产环境 | 实操依赖审查减少 95%;安全工程师每周节省 >5 小时 | 供应商案例研究;未披露合同范围或续约数据 |
| Replit | AI 编码 / 开发者平台 | 将 GitHub check 接入依赖审查工作流,筛查新增和传递依赖包 | 生产环境 | 误报定性减少,发版时更快建立信心 | 未披露量化时间或预算成效 |
| Vercel | 开发者云 / monorepo 平台 | monorepo 依赖卫生工作流,支持 pnpm,并分阶段推广 | 生产环境 | 依赖蔓延、认知负担和手工包评估减少 | 没有公开节省指标或席位数 |
| Cedar | 医疗金融软件 | 面向精简安全团队的 GitHub 原生可达性与漏洞分诊 | 生产环境 | 警报减少 70%;工作量从每月大约 30-40 张工单降至 10-12 条警报 | 成效来自 Socket 案例研究,并非独立审计 |
| Chia | 开源区块链平台 | 围绕 GitHub 的审查流程,服务大型公开代码库和贡献者社区 | 生产环境 | 未处理安全警报下降 70%;工程师在 GitHub 内处理 90% 任务 | 开源工作流未必能外推到所有企业买家 |
| JumpCloud | 身份 / 合规敏感 SaaS | 覆盖 600+ 个仓库的全 repo 推广,包含可达性、许可证、SBOM 和开发者端点覆盖 | 生产环境 | 50 个团队马上获得可见性,手工库审查负担下降 | 未披露商业合同或扩张金额 |
| Render | 云基础设施 | 以 PR 评论切入,配合许可证扫描和后台包审查 | 生产环境 | 多年留在 PR 流程中,因为噪音一直低 | 留存代理指标是定性的,不基于收入 |
| SHI | 企业技术解决方案 | GitHub app 加浏览器扩展辅助的包研究,服务小型专业团队 | 生产环境 | 节省数百个工程师小时;估计 ROI 为 400-500% | 内部产品组证据比全客户群证据更窄 |
该枚举只覆盖经审查公开证据中最强的一部分。Socket 还列出其他客户,但若干 logo 尚无公开部署细节或成效证据。
[CU002, CU009, CU010, CU012, CU014, CU015]公开证据质量差异很大:既有量化案例,也有几乎没有部署深度披露的标杆客户标识。
评级是对公开证据深度的定性评估。强表示具名部署加具体工作流或结果细节;弱表示仅有客户标识提及,部署细节很少或没有。
[CU002, CU009, CU010, CU012, CU014, CU015]6.3 留存代理、客户声音与运营摩擦
Socket 没有公开披露 NRR、GRR、合同期限、客户留存队列,或从免费 / 开源使用转化为付费企业账号的客户数 [CU026, CU035]。 因此,公开耐久性只能从工作流粘性、评测情绪和更深运营嵌入证据中推断。这些代理指标方向上正面。Render 称 Socket 多年来一直留在 PR 中,因为它保持低噪声;JumpCloud 把它输入覆盖 50 个团队和 600+ 个代码库的内部评分;JupiterOne 把它视作 CI/CD 策略门;Replit 和 Doctolib 则把它连接到合规证据和客户保证 [CU017, CU018, CU019, CU040]。 独立评测也大多正面,尽管严谨度不如具名案例研究。AppSecSanta、ToolRadar、Startupik 和 MakerStack 都把 Socket 描述为靠行为分析、 PR 级反馈和主动供应链检测区别于经典 CVE 匹配 [CU027]。与此同时,这些评测也持续指出,平台仍在成熟, 在 npm 或 JavaScript 为主的环境中最强,并且最好与传统 CVE 扫描器搭配使用,而不是完全替代 [CU028, CU038]。 反向证据也重要。2025 年 1 月一篇 Medium 测试称,Java 依赖没有出现在 Socket 的 UI 或 PR 评论中,即使支持团队承认并部分修复了已报告问题; 作者认为,如果买方假设生态覆盖是通用的,这可能制造虚假的安全感 [CU029]。Socket 自己的 Vanta 集成文档也指出,OAuth 令牌经常被撤销, 这会为合规用户制造同步缺口,直到连接重新授权 [CU030]。这些问题并不抵消强 GitHub 原生客户背书,但确实限定了留存故事: 当 GitHub、JavaScript/Python 和低摩擦开发者工作流是买方环境核心时,产品最容易承保 [CU038]。
| 指标 / 代理信号 | 数值 | 分群 | 置信度 | 尽调问题 |
|---|---|---|---|---|
| 独立汇总评分 | ToolRadar 称各评论平台汇总为 4.6/5 | 潜在买家 / 广泛市场 | 低 | 获取该汇总背后的原始评论平台组合和已验证评论者数量 |
| 独立判断 | MakerStack 给 Socket 7.4/10,并称最适合 npm/PyPI 使用重的团队 | 技术评估者 | 低 | 核查非 JavaScript 团队推广后是否报告类似价值 |
| 工作流长期留存代理信号 | Render 称 Socket 多年来一直留在 PR 中,因为噪音保持较低 | 云 / 开发者平台买家 | 中 | 索取 logo 级续约数据或合同历史,验证该代理信号 |
| 运营嵌入代理信号 | JumpCloud 将 Socket 接入 50 个团队、600+ 个仓库的内部评分 | 身份 / 合规买家 | 中 | 索取付费扩张指标,并确认使用是否在全球标准化 |
| 合规嵌入代理信号 | Replit、Doctolib 和 JupiterOne 描述了 Vanta、审计或 CI 策略集成 | AI 编码 / 受监管 SaaS / 安全买家 | 中 | 询问这些合规集成多常影响续约或增购 |
| 正式留存经济性 | 未公开披露:NRR、GRR、总流失率、合同期限、续约队列 | 全部分群 | 高 | 索取队列数据、logo 留存、美元留存和合同期限分布 |
公开留存证据大量依赖代理信号。评分和工作流嵌入信号有帮助,但不能替代已披露的续约或收入留存队列。
[CU011, CU017, CU018, CU019, CU027, CU028]| 来源 | 信号类型 | 观察 | 客户影响 | 反向观点 / 待追问 |
|---|---|---|---|---|
| Medium(Jan 2025 测试) | 反向产品质量信号 | 测试者报告 Java 依赖即使在支持团队承认一个 bug 后,仍未出现在 UI 或 PR 评论中 | 假设广泛生态覆盖的买家,可能产生虚假的安全感 | 需要对 Java 和其他非核心生态做更新后的独立验证 |
| AppSecSanta | 独立评测 | 肯定行为检测和 GitHub PR 集成,但称 Socket 应补充而非替代传统 CVE 扫描器 | 支持其作为前置防御层的定位,而不是完整平台替代 | 询问生产环境客户多常把 Socket 与另一款 SCA 工具搭配使用 |
| ToolRadar | 评论聚合方 | 突出 4.6/5 汇总信号,但指出平台较新、聚焦 npm、存在学习曲线,以及付费企业功能 | 暗示更适合成长中的技术团队,而不是所有企业买家 | 需要 npm-heavy 组织之外的已验证客户背书 |
| MakerStack | 独立分析师式评测 | 给产品 7.4/10,并称 JavaScript 仍是最强生态,且没有 self-hosted 选项 | 可能放缓受监管或非云买家的采用 | 索取 Java、Ruby、Rust 及 self-hosted 敏感账号中的路线图和附加率 |
| Socket Vanta 文档 | 供应商文档中的注意事项 | OAuth token 经常被撤销,可能让合规同步看起来失效,直到重新授权 | 以合规为主导的买家即使部署后也可能遇到运营摩擦 | 索取 token 撤销频率和缓解路线图 |
本表刻意混合正面和怀疑性的外部视角,避免本章只依赖供应商挑选的客户故事。
[CU027, CU028, CU029, CU030, CU038]6.4 扩张循环与集中度风险
公开资料里,Socket 最可信的扩张路径不是披露更多席位数,而是低门槛初始落地之后继续扩产品面。案例显示了反复出现的路径:客户先从 GitHub App 或 PR 阶段依赖扫描起步,再延伸到 API 审批、可达性分析、许可证和 SBOM 工作流、CI/CD 卡点、Vanta 同步、依赖搜索或开发者端点防护 [CU007, CU031, CU041]。这条路径把一个战术型依赖扫描器,变成更大治理与合规栈的一部分。所以 JumpCloud、JupiterOne、Replit、Doctolib、MetaMask 等客户谈的是 SBOM、CI 策略、历史依赖搜索、审计证据和开发者机器防护,而不只是一次性告警 [CU011, CU016, CU018, CU021, CU031]。 集中度风险在于,这套扩张逻辑仍建立在偏窄的公开客户画像上。Socket 最知名的公开背书仍集中在 AI 实验室、开发者工具、开源占比较高的平台、合规厂商以及云 / 安全团队 [CU034, CU037]。2026 年,AI 驱动开发正在放大依赖风险,这是一组有吸引力的客户;但也意味着公开客户品牌集中在已经很成熟、而且往往以 GitHub 为中心的组织 [CU037, CU038]。评论和反向 Java 测试从另一个角度强化了同一点:Socket 在 npm/Python 和 GitHub 中心工作流里看起来最强,放到更复杂的企业异质环境,公开证据还没充分跑通 [CU028, CU029, CU038]。 外部佐证能帮 Socket 提升认知度,但不能证明经济性。新闻报道大多复述 Socket 的客户名单和使用指标,而不是披露采购细节、留存队列或收入集中度 [CU039]。因此,公开尽调仍需要管理层回答付费客户结构、头部客户暴露,以及 AI 原生 logo 组合究竟是代表性客户基础,还是精心挑选的领导客户切入口 [CU036, CU039]。
| 扩张驱动因素 | 集中度 / 执行风险 | 影响 | 尽调路径 |
|---|---|---|---|
| GitHub App 或 PR check 首次部署 | 高度依赖以 GitHub 为中心的工作流,可能限制其在异构 SCM 环境中的适配 | 获客切口强,但存在平台集中风险 | 按 SCM、包生态和部署模式拆分收入 |
| API、可达性与 CI/CD 策略扩张 | 如果治理深度不是刚需,买家可能停留在战术扫描 | 增购取决于能否证明价值不止降低噪音警报 | 索取按模块和客户分群的附加率 |
| 合规工作流:SBOM、Vanta、审计、许可证策略 | Vanta token 撤销等集成脆弱性会削弱信任 | 可切入合规预算的交叉销售动作 | 索取活跃集成使用、故障率和扩张胜率 |
| AI-native 与开发者工具客户密度 | 公开品牌集中在热门但相对狭窄的买家群 | 品类领导力是正向信号,但若 AI-native 团队转向平台原生工具,存在风险 | 索取 AI-native logo 和前十大客户的收入占比 |
| 头部账号 logo 带来的认知度 | 在已审查公开证据中,xAI、Cursor、Figma、Gusto 和 Mercado Libre 仍只是 logo 证据 | 若不与实测部署分开,可能夸大生产深度 | 索取头部 logo 的部署阶段、合同规模和可背书性 |
| 广泛平台覆盖指标 | 若免费或事件驱动使用占比高,27,000 个受保护组织可能夸大商业深度 | 可能掩盖集中度或较弱的免费转付费 | 索取付费客户数、免费转付费转换率和分群 ARR 组合 |
此处扩张指首次部署后的产品覆盖深度,不是已披露的净收入留存。集中度风险是公开记录缺口,需要管理层披露,而非靠假设补齐。
[CU023, CU024, CU026, CU030, CU031, CU036]公开扩张路径从 PR 时扫描开始,再深化到策略、合规和端点覆盖。
这是定性的部署流程,不是量化转化漏斗。公开资料没有披露赢单率或各阶段数量。
[CU007, CU011, CU031, CU032, CU033, CU041]6.5 图表
07风险
7.1 风险概览与严重程度排序
Socket 最实质的风险集中在差异化能否守住,而不是公司能不能活下去。公司刚拿到新资金,也有看得见的客户证据;但核心投资问题是,Socket 的检测和优先级排序能否持续明显好过大型平台逐步打包进现有开发者工作流的功能。GitHub 现在已经提供 dependency review、Dependabot、malware alerts、SBOM 支持、advisory data 和 artifact attestations,Snyk 也有自己的可达性分析。Socket 的回应是更宽的恶意包检测、Coana 带来的可达性分析和人工验证分析;但这个回应也抬高了执行门槛:客户一旦为了降低噪音而购买 Socket,精准度回退或工作流摩擦上升,都会很快削弱付费意愿。公开可见的最强缓释因素,是公司的不上传源代码模型、透明状态页、企业控制项,以及能降低近期融资压力的 Series C。公开证据最弱的地方仍是可持续性:客户材料展示了广度和案例胜利,却没有展示留存、集中度或队列经济性。[CR001, CR003, CR020, CR023, CR028, CR036]
以可能性和影响力两维呈现 Socket 截至 2026-05-24 的主要风险。
影响力和可能性分桶是分析师定性判断,锚定产品、市场和披露证据,而不是管理层给出的概率。
[CR020, CR023, CR030, CR036, CR045, CR046]7.2 竞争与平台依赖风险
Socket 在战略上缠在一批也会压缩品类的平台上。它自己的入门指南把 Socket for GitHub 称为最容易、也最强的部署路径;从获客角度这很合理,但风险在于 GitHub 同时控制 PR 界面、安全公告图谱、dependency review,以及塑造买方预期的捆绑安全功能。Dependabot 和 dependency review 已经覆盖许多工程组织在已知漏洞策略执行上的一大块需求,GitHub 还能把这些控制项变成工作流默认预期。Snyk 的可达性工作削弱了另一个历史差异点:Coana 之后,可达性比以往更重要,却不再独特。与此同时,npm 和 GitHub 正靠可信发布、来源证明和 attestations 持续抬高安全基线。这些变化不会消灭恶意包风险,但会让 Socket 的长期护城河更依赖精准度、策略能力、响应速度和企业工作流适配。Socket 有机会在 GitHub 之外扩企业客户,但定价和文档显示,很多广度——GitLab、Azure DevOps、自托管仓库、SCIM、审计日志和 IP 控制——都留给了更高阶套餐。[CR020, CR021, CR022, CR023, CR024, CR025]
| 依赖 | 交易对手 / 平台 | 角色 | 集中度 | 失效场景 | 严重性 | 缓解措施 | 剩余暴露 |
|---|---|---|---|---|---|---|---|
| 原生代码安全与供应链功能 | GitHub / Microsoft | 在 PR 工作流、安全公告数据、依赖审查、恶意软件警报和证明上竞争 | 高 | GitHub 扩大内建覆盖,足以让买家把 Socket 视为可选附加组件 | 极高 | Socket 以恶意包行为分析、企业策略和更深可达性区分 | GitHub 仍控制主要工作流表面,并可能压缩品类定价 |
| 以 GitHub 为中心的获客与部署路径 | GitHub | 分发和最容易的公开部署路径 | 高 | GitHub API、checks 体验或买家偏好变化,削弱 Socket 最容易的采用路径 | 高 | Enterprise 层级加入 GitLab/Azure/自托管选项和 CLI/Firewall 替代路径 | 公开材料仍把 GitHub 呈现为默认且最强路径 |
| 可达性分析竞争 | Snyk | 争夺优先级排序和漏洞噪音降低预算 | 中高 | Snyk 补足足够多的精度缺口,使 Coana 不再让 Socket 形成实质差异 | 高 | Socket 营销全应用可达性、恶意包检测和更广的供应链信号 | 可达性不再独特,因此买家可直接按工作流适配和噪音成效做基准比较 |
| 注册表生态的安全基线改善 | npm / GitHub / CI 提供商 | 可信发布、来源证明和工件证明降低一类供应链滥用 | 中 | 更多买家预算转向现有平台已提供的基线控制 | 中高 | Socket 仍处理来源证明单独无法解决的恶意行为、策略和响应工作流 | 基线控制仍可能收窄感知差异化和定价权 |
| 核心服务提供商 | AWS S3、Render、Stripe、WorkOS 与 Vanta | 存储、托管、支付、身份和合规同步 | 中 | 宕机、token 频繁变动或政策变化削弱产品交付或企业采购信心 | 中高 | Socket 使用主流提供商,并公开暴露部分状态 / 合规工具 | 提供商集中度仍有意义,因为若干控制不由第一方拥有 |
| 非 GitHub 企业扩张 | GitLab、Azure DevOps、Bitbucket、自托管 SCMs | 用于覆盖 GitHub 之外更广的企业 | 中 | 扩张较慢,因为这些路径锁在更高层级,且设置更复杂 | 中 | Enterprise 层级已记录支持和集成 | 公开自助式动能仍看起来在 GitHub 最强 |
各行聚焦 Socket 外部依赖,这些依赖可能压缩增长、削弱精度优势或扰乱交付。
[CR020, CR021, CR022, CR023, CR025, CR026]展示平台竞争、精度风险和披露缺口如何传导到收入质量与估值信心。
传导路径代表分析上的因果关系,并非管理层披露的内部评分卡。
[CR020, CR023, CR028, CR036, CR040, CR046]梳理影响 Socket 产品交付和企业销售动作的关键外部平台与提供商。
本图强调会实质影响采用、可靠性或合规姿态的外部依赖。
[CR016, CR018, CR032, CR033, CR035, CR038]7.3 产品质量、覆盖与运营风险
Socket 的产品承诺是在真实开发者约束下保持精准,因此即便收购 Coana 之后,误报和漏报风险仍是核心。Coana 的可达性引擎和 Socket 的定价主张都指向大幅降噪;但正是这些主张,会让动态环境或覆盖不完整环境里的精准度下滑变得更昂贵。Socket 的不上传源代码架构是采购优势,却也带来盲区:公司自己的 known-issues 页面称,私有 npm packages 会被跳过,除非客户另行启用私有仓库,或把 package 重构为 workspace。不同生态和界面的覆盖也不均衡。公开矩阵里,GitHub Actions 缺少可达性和自动修复;Swift 仍停留在 CVE-only;若干生态仍处于计划中或不支持。Socket 加入浏览器扩展、AI 模型扫描、GitHub Actions 分析和更多语言引擎后,运营复杂度还在升高。状态页展示了多少离散组件必须保持健康,CI/CD 集成指南也显示,客户仍要正确管理 API keys、tokens 和 branch-protection 细节。Vanta 集成进一步说明,当第三方 token 行为不稳定时,合规自动化会在运营上变脆。[CR004, CR005, CR006, CR007, CR008, CR010]
| 失效模式 | 可能性 | 严重性 | 缓解成熟度 | 剩余暴露 | 未解决缺口 |
|---|---|---|---|---|---|
| Coana 可达性扩展到更多生态和表面后,警报精度回退 | 中高 | 高 | 中 — Coana、定价声明、人工验证和策略控制已经存在 | 如果客户仍看到高噪音,核心价值主张和付费意愿会很快削弱 | 没有公开基准集显示按生态、客户规模或警报类别维持长期精度 |
| 仅看 manifest 与私有包盲点带来的漏报 | 中 | 高 | 中 — 无源代码模型改善隐私和采购 | 除非客户调整 repo 设置,私有 npm 包和间接代码路径可能仍分析不足 | 需要客户证据说明私有包或动态语言边缘案例多常逃过早期检测 |
| 语言生态成熟度不均、跨语言功能只部分对齐 | 高 | 中高 | 中 — 路线图和 beta 标签明确 | 异构技术栈客户在核心生态之外可能只获得 CVE-only 或降级价值覆盖 | 需要按生态成熟度层级拆分附加率与流失 / 赢单-输单数据 |
| API、仪表盘、状态跟踪分析服务、AWS S3 或 Render 出现中央服务宕机或提供商故障 | 中 | 中高 | 中 — 有公开状态页和标准云态势 | 多组件故障可能同时扰乱阻断、扫描、仪表盘或报告检索 | 需要内部 RTO/RPO、事件历史和基础设施提供商集中度细节 |
| API key、受保护变量、刷新 token 和分支策略配置导致 CI/CD 与合规集成脆弱 | 中 | 中 | 中 — 文档提供设置指导和企业治理控制 | 误配置或 token 频繁变动可能悄悄降低覆盖,或让客户觉得产品不可靠 | 需要按集成拆分的支持工单数据,以及 GitHub/GitLab/Vanta 问题解决时长证据 |
| 向 GitHub Actions、AI 模型、扩展和多语言引擎延伸带来的路线图与支持复杂度 | 高 | 中高 | 低中 — 状态透明度和定价分层有帮助 | 广度会在一家约 100 人公司内稀释 QA、支持和研究焦点 | 需要产品线人员配置图、发布节奏稳定性,以及按表面拆分的上线后缺陷趋势 |
运营风险按其削弱客户信任、采购成功或续约结果的直接程度排序。
[CR006, CR007, CR010, CR012, CR013, CR015]7.4 法律、监管与信任风险
Socket 公开做了几件隐私友好的事——最重要的是它说自己从不上传源代码——但相对于当前产品面,它的法律和合规披露仍显得偏薄。尽管 2025-2026 年公司扩展到更多集成、AI 场景界面、企业控制项和更广分析类别,隐私政策仍停在 February 2022。政策也明确涵盖第三方处理,以及按法律或政府要求披露,因此跨境传输机制会影响企业销售。EU-US Data Privacy Framework 有帮助,但当受监管客户要求最新的传输、子处理方和数据处理细节时,它不能替代保障措施、尽调和更新文档。EU Cyber Resilience Act 进一步抬高前瞻负担:2026 年开始引入报告义务,2027 年扩展到更广生命周期义务。合同层面,Socket 的公开 agreements 页面证明协议存在且为当前版本,但没有说明营销摘要之外的责任上限、赔偿范围或 uptime 承诺究竟如何。公开执法搜索在本次审查中没有发现针对 Socket 的 FTC 事项,但这只是一个监测信号,不等于其他地方不存在潜在索赔、客户争议或监管问题。[CR029, CR030, CR031, CR041, CR042, CR043]
| 规则 / 合同 / 暴露 | 司法辖区 | 状态 | 可能性 | 严重性 | 缓解措施 | 剩余暴露 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| GDPR / EU-US 传输合规与过时隐私披露 | EU / US | 隐私政策公开,但最后更新于 2022-02-07;传输依赖第三方提供商和法律程序例外条款 | 中 | 高 | 无源代码模型减少数据量;DPF 和其他保障措施覆盖 US 传输 | 企业买家在购买或续约前,可能仍要求更新的 DPA、子处理商和 EU 专项控制 | 索取当前 DPA、子处理商列表、保留期限表,以及法律 / 隐私文档已按 2026 产品范围更新的证据 |
| Cyber Resilience Act 软件生命周期义务 | European Union | CRA 已生效;报告义务从 2026-09-11 开始,主要义务从 2027-12-11 开始 | 中 | 中高 | Socket 已对外宣传漏洞处理、企业控制和安全态势 | 生命周期、报告和证据生成义务仍可能增加合规成本和产品流程负担 | 向管理层索取 CRA 准备计划、报告责任人,以及为 EU 客户生成产品安全证据的方式 |
| 责任、赔偿和 SLA 条款的公开合同不透明 | US / 全球企业合同 | 协议页面为最新,但公开抓取未揭示实质性企业责任条款 | 中 | 中高 | 当前协议版本可见,定价页面宣传 Enterprise 的 uptime SLA | 仅凭公开材料,投资人无法承销保修、赔偿、数据处理或服务补偿暴露 | 获取当前企业 MSA、EULA、DPA 和 SLA 附表,包括责任上限、赔偿、安全承诺和排除条款 |
| 公开执法 / 争议可见性缺口 | US 与非 US | 本次审查未发现 Socket 相关 FTC 事项,但公开数据库核查并不穷尽 | 低中 | 中 | 未发现公开执法信号;存在公开监测渠道 | 被威胁的索赔、私人争议或非 US 事项仍可能存在于已检索公开表面之外 | 索取完整诉讼清单、被威胁索赔日志、客户争议摘要,以及过去 24 个月的任何监管问询往来 |
行顺序按当前剩余投资相关性排序,而非任何官方公司风险分类法。
[CR029, CR030, CR031, CR041, CR042, CR043]7.5 人员、执行与证据缺口
Socket 的公开势头是真实的:Series C、具名客户和广泛案例库足以证明市场牵引力。但证明市场被需要,不等于证明可持续。公开材料展示了许多 logo 层面和案例层面的胜利,尤其集中在减少告警和提升工作流效率;但在 $1 billion 估值轮之后,投资人通常需要的年经常性收入(ARR)、净留存率(NRR)、流失率或集中度,公开材料并没有给到足够具体的披露。广度放大了执行风险。SecurityWeek 报道 Series C 时 Socket 大约有 100 名员工,而 Socket 同时在营销或文档中覆盖 GitHub、GitLab、Azure DevOps、AI 模型扫描、浏览器扩展、firewall、certified patches、可达性、企业治理等更多方向。创始人驱动的可信度是优势——Feross Aboukhadijeh 和顾问网络显然有助于信任与招聘——但这种公开定位也集中了承载声誉和产品原创叙事的风险。收购 Coana 缓解了一部分技术缺口,却也带来对新整合、研究属性强团队的留任依赖,同时还要把这些专长转成可复制的企业产品结果。[CR001, CR002, CR004, CR005, CR036, CR037]
| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重性 | 缓释措施 | 尽调路径 |
|---|---|---|---|---|---|
| 创始人背书的产品可信度 | Feross 是公共安全可信度、产品叙事和开发者信任信号的核心 | 中 | 高 | 强顾问和投资人阵容部分降低单点声誉风险 | 索取继任覆盖方案、高级产品 / 安全领导层地图,以及大客户关系归属 |
| Coana 技术团队留任 | 刚收购的研究型团队现在支撑精准度叙事 | 中 | 高 | 整个团队已加入,整合很快启动 | 复核留任方案、产品归属,以及是否依赖 Aarhus 专家中的关键个人 |
| 执行带宽与业务宽度 | 约 100 名员工,要同时扩产品、集成和生态 | 高 | 高 | 新 Series C 轮资金支撑招聘和路线图落地 | 索取按产品线拆分的组织图、支持负载,以及主要界面的缺陷积压 |
| 客户证据耐久性缺口 | 案例证明覆盖面和工作流胜出,但不能证明留存、集中度或合同条款 | 高 | 中高 | 已点名客户和案例表明采用信号有分量 | 要求提供前 20 大客户集中度、NRR/GRR、logo 流失、合同期限和扩张队列 |
这里的执行风险不主要来自生存层面的现金压力,而是 Socket 必须同时放大精准度、平台宽度和企业客户证据。
[CR001, CR004, CR005, CR037, CR039, CR040]7.6 缓释因素、监测指标与叫停条件
狭义运营层面,Socket 看起来并不脆弱:它有新资金、透明状态页、不上传源代码架构、企业身份控制项,以及明确的分诊机制来压低噪音。这些都是有用的缓释因素,但它们不等于品类结构仍有利的证据。因此,投资人应监测少数可度量的叫停条件,而不是把所有风险等量看待。第一是竞争压缩:如果 GitHub 持续扩展默认依赖、恶意软件和策略覆盖,而 Socket 的胜利故事仍以 GitHub 为中心,续约压力会在增长失速上新闻前先出现。第二是告警可信度:Socket 自己的叙事围绕更少噪音和更可执行的优先级排序,公开或客户证据一旦显示误报疲劳上升,就会打到投资逻辑的核心。第三,Coana 整合必须转化为稳定的企业产品表现,同时不流失关键技术团队。第四,法律 / 合规材料需要追上 2026 年产品面。最后,下一次融资或重大客户里程碑,应该带来比今天公开市场更好的可持续性披露。[CR007, CR018, CR024, CR028, CR030, CR033]
| 风险 | 可监控触发项 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| GitHub 平台挤压 | GitHub 默认推出显著更宽的依赖、恶意软件或策略控制能力 | 续约时,主要买家群体可以用 GHAS/Dependabot 加原生策略检查替代 Socket | 除非 Socket 证明精准度更强且胜单可留住,否则将投资判断从高溢价增长切到价格压力情景 |
| 告警精准度恶化 | 公开案例或客户背书不再提显著降噪,或支持负担急剧上升 | 任何证据显示,动态或多语言环境下的误报降低主张不再成立 | 视为打破投资假设,因为精准度是产品核心承诺 |
| Coana 整合延误 | Coana 核心负责人离开,或可达性功能在重要生态的推出停滞 | 功能广泛嵌入前,企业推出里程碑未达成或团队成员离职 | 提高执行折扣,并要求可达性分析带来持续客户价值的证据 |
| 法律 / 隐私披露滞后 | 企业扩张之下,仍没有更新 DPA、子处理方清单或刷新后的隐私 / 法律材料 | 再过 2-3 个季度,公开文档或面向客户的证据包仍未刷新 | 在押注受监管客户增长前,假设采购摩擦存在,并要求管理层深度尽调 |
| 可用性 / 集成脆弱性 | GitHub、GitLab 或 Vanta 流程反复受到宕机或 token 集成故障影响 | 关键集成出现重大事故频率,或支持升级长期持续 | 下调扩张假设,把运营可靠性视为董事会层面问题 |
| Series C 轮后的耐久性证据缺口 | 下一轮重大融资或董事会材料仍缺少 ARR、NRR、流失或集中度细节 | 下一轮估值上调融资或重大老股交易前,公开材料或尽调室中的耐久性指标没有改善 | 没有留存和集中度质量的私下证据,就不要给高倍数定价 |
这些触发项按可观测性选择:每一项都可以外部监控,或在尽调中明确索取。
[CR018, CR028, CR030, CR033, CR036, CR046]7.7 图表
08估值
8.1 融资背景与规模证据
Socket 的 May 2026 Series C 描述起来容易,写进投资模型更难。可观察部分很强:公司称自己以 $1 billion 估值融资 $60 million,使累计融资达到大约 $125 million,并把 Thrive Capital、a16z、Abstract Ventures 和 Capital One Ventures 加进了一个高质量投资团。公开运营证据也强过典型 Series C。Socket 披露了 Anthropic、xAI、Figma、Vercel、Mercado Libre 等有辨识度的客户;官网称保护了 27,000+ 个组织、300,000+ 个仓库、1.5 million 名受信任开发者、每月 11.6 million 个受保护 commits,并且每周拦截超过 10,000 次攻击。2025 年收购 Coana 又提供了第二个重要信号:Socket 称收入较前一年增长超过三倍,并借这笔交易把可达性分析带入平台,意图降低告警噪音,把产品从基础 SCA 往外推。AI 编码扩大第三方代码进入生产的规模之后,投资人愿意为软件供应链安全里的品类领导者支付溢价,这些都是可信理由。 仍然难以承销的部分在于透明度。公开来源没有披露年经常性收入(ARR)、净留存、毛利率、现金消耗,或决定 $1 billion 企业价值对新资金是否真正有吸引力的股权结构瀑布。因此,本轮价格可以作为战略品类押注来辩护,但还没有完全证明是一笔由基本面支撑的便宜交易。[CV001, CV002, CV003, CV004, CV005, CV006]
| 维度 | 评估 | 信心 | 决策含义 |
|---|---|---|---|
| 建议 | 跟踪 / 继续研究;仅保留有条件兴趣 | 中 | 不要把公司质量等同于 2026 年 5 月这一轮便宜的证据。 |
| 风险评级 | 高 | 中 | 业务质量可见,但变现和股权结构不确定性仍可能快速压缩价值。 |
| 估值立场 | 仅在 ARR 达到中双位数百万美元时才算合理;按公开证据略偏贵 | 中 | 高于本轮价格或上调乐观估值前,先要求经济性证据。 |
| 证据质量 | 有所改善但仍不完整 | 中 | 公开证据对产品、客户和增长信号的覆盖好于单位经济。 |
| 决策含义 | 把 $1B 当尽调锚点,不要当出清价 | 中 | 本轮可以辩护,但当前 ARR / NRR / 烧钱披露决定它是否真正可投。 |
这条建议明确对价格敏感:Socket 看起来有价值,但 $1B 这一轮仍需要经济性验证。
[CV001, CV003, CV049, CV056, CV057, CV058]| 维度 | 正向假设 | 反向假设 | 哪些证据会改变判断 |
|---|---|---|---|
| 品类时点 | AI 编程和上升的依赖包风险,把软件供应链安全推到企业更高优先级。 | 品类紧迫性不自动转化为持久付费席位增长。 | 展示与 AI 驱动采用相绑定的企业转化、续约和扩张数据。 |
| 产品差异化 | Coana 可达性分析、Firewall,以及 2025-2026 年快速功能扩展,支撑高溢价平台叙事。 | 如果高端模块不能显著抬高 ACV,买家仍可能把 Socket 看成单点方案。 | 披露可达性分析、补丁和 Firewall 的挂载率及 ACV 提升。 |
| 商业证明 | 头部客户 logo 和 27,000+ 个受保护组织,暗示真实市场拉力。 | 免费 / 开源使用可能把覆盖面做大,却不对应付费收入。 | 按套餐层级拆分受保护用户、付费席位和企业 ACV。 |
| 可比公司支撑 | JFrog、PANW 和 Wiz 显示,市场愿意为强开发者安全平台支付高倍数。 | GitLab 和 SentinelOne 显示,公开市场也会很快重估到低得多的收入倍数。 | 提供当前 ARR、NRR、毛利和增长,把 Socket 可信地放进可比区间。 |
| 竞争 | Socket 的开发者优先工作流和威胁研究节奏,形成了真实产品身份。 | GitHub 把 AppSec 打包进原生工作流,Copilot 正变成 AI 时代的开发者控制平面。 | 展示相对 GHAS 的胜率,以及大客户中长期工作流所有权的证据。 |
反向假设核心是变现和打包压力,而不是否定需求。
[CV004, CV012, CV015, CV016, CV020, CV022]可观察证据真实存在,但经济性缺口仍让轮次价格下的买入结论不够干净。
流程把定性的 IC 决策链压缩为六个节点。
[CV001, CV006, CV012, CV045, CV049, CV057]8.2 可比公司集合与收入代理
公开可比公司给不出一个干净答案,但它们界定了 Socket 在 $1 billion 定价之下必须跑出的经济性。GitLab 的收入倍数约 4.7x,SentinelOne 约 6.4x;若按这个区间,Socket 需要 ARR 已经大约达到 $150 million-plus,才能支撑本轮。JFrog 是更相关的软件供应链和 DevSecOps 平台,按 FY2026 收入指引的倍数更接近 14x,意味着 Socket 大约需要 $70 million ARR。Palo Alto Networks、CrowdStrike 等高溢价 cyber 平台,以及 Wiz 等高溢价私有可比公司,可以支持高得多的倍数,但前提是规模大得多,或披露收入清楚得多。Chainguard 显示了市场愿意给高增长软件供应链安全多高定价,但它披露的倍数是离群值,不是中位数。 Socket 自己的公开足迹能拼出一个粗略代理。公司披露 1.5 million 受保护开发者,Team 和 Business 的标价分别是每名开发者每月 $25 和 $50;若按 $35 混合价格,每 1% 开发者变现大约对应 $6.3 million ARR。这意味着,要用 20x 倍数支撑 $1 billion 估值,需要披露开发者基数约 8% 变现,或者规模更小的企业客户群通过可达性、firewall、certified patches 和更广企业控制项支付高得多的有效 ARPU。由于 Socket 对开源免费,并且有 $0 入门层,受保护用户数是噪音很大的收入代理,而不是收入披露。所以这座可比桥只能说明「可能」,不能说明「已证明」。[CV022, CV023, CV024, CV025, CV026, CV027]
| 可比公司 | 状态 | 收入 / ARR 锚点 | 估值 / 市值 | 隐含倍数 | 参照意义 | 局限 |
|---|---|---|---|---|---|---|
| GitLab | 上市 DevSecOps 平台 | $955.2M FY2026 收入 | $4.51B 市值 | ~4.7x | 为内嵌安全能力的开发者工具提供上市公司下限 | 公开市场倍数反映增长较慢、产品范围更宽。 |
| SentinelOne | 上市网络安全平台 | $1.001B FY2026 收入 | $6.38B 市值 | ~6.4x | 规模化但仍亏损的安全软件上市中档参照 | 更接近终端安全,不是供应链开发者工作流。 |
| JFrog | 上市软件供应链 / DevSecOps | $628M-$632M FY2026 收入指引 | $8.96B 市值 | ~14.2x | 保护软件交付链的最贴近上市可比公司 | 用的是前瞻指引,不是已完成财年。 |
| Palo Alto Networks | 规模化上市网络安全平台 | $9.2B FY2025 收入 | $211.33B 市值 | ~23.0x | 表明投资人愿为分发强、覆盖广的安全平台支付溢价 | 规模、分散度和成熟度都远高于 Socket。 |
| Wiz | 私营高溢价云安全可比公司 | ~$350M ARR (2024) | $12B 估值 | ~34.3x | 快速扩张安全龙头的私营溢价上限参照 | 产品范围不同,披露 ARR 基数也大得多。 |
| Chainguard | 私营软件供应链可比公司 | $40M ARR,近期目标 >$100M | $3.5B 估值 | ~87.5x 当前 / ~35x 目标 | 最接近的软件供应链高增长估值参照 | 倍数属离群值,且是 2025 年披露,不是稳定中位可比。 |
可比估值计算旨在圈定 Socket 需要多少 ARR 才能支撑 $1B,而不是暗示存在一个完美对标。
[CV027, CV030, CV033, CV036, CV039, CV041]$1B 估值需要多少 ARR,取决于投资人认为 Socket 应拿到哪一档倍数,差异很大。
敏感性同时使用公开可比公司倍数和一个已披露覆盖面的变现代理值;它只是示意,不是管理层预测。
[CV027, CV030, CV033, CV036, CV039, CV045]8.3 乐观、基准与悲观估值区间
情景框架应该保持简单,因为最大未知数是当前 ARR。悲观情景下,Socket 仍是一家真实公司,技术也强,但公开规模信号转化为收入的速度慢于本轮隐含预期。这大致对应 $25 million to $35 million ARR 和十几倍中段到 20x 的倍数,得到大约 $450 million to $700 million 价值。基准情景下,Socket 已经把受保护开发者足迹中的一个有意义少数,成功转成付费席位、企业套餐,以及 ARPU 更高的可达性或 firewall 模块。这能支撑大约 $45 million to $60 million ARR 和约 18x to 22x 倍数,对应 $800 million to $1.1 billion 附近。 乐观情景需要的不只是好 logo 和快速产品节奏。它需要证据证明,2025 年收入增长超过三倍已经转化为 2026 年持续规模化,变现更接近高个位数开发者转化率或等效企业 ACV,而且市场继续把软件供应链领导者当作 Wiz 或 Chainguard,而不是增速更慢的公开 DevSecOps 公司。在这些条件下,$65 million to $85 million ARR 和 20x to 25x 倍数可以支撑大约 $1.2 billion to $1.7 billion。概率加权中枢仍略低于本轮,因此本轮看起来可辩护,但并不显然便宜。[CV045, CV046, CV047, CV048, CV049, CV050]
| 情景 | 概率 | ARR 代理假设 | 倍数区间 | 估值区间 | 主信号 |
|---|---|---|---|---|---|
| 悲观 | 30% | ~$25M-$35M ARR | 15x-20x | $450M-$700M | 受保护覆盖面转化缓慢,打包平台压住变现。 |
| 基准 | 45% | ~$45M-$60M ARR | 18x-22x | $800M-$1.1B | Socket 将覆盖面中有意义的少数转化为高端企业收入。 |
| 乐观 | 25% | ~$65M-$85M ARR | 20x-25x | $1.2B-$1.7B | 收入继续三倍增长,企业 ARPU 抬升,市场把 Socket 视为高溢价私营可比公司。 |
| 概率加权视图 | 100% | 加权中心约在 $800M 高段到 $900M 低段 | 混合 | $0.88B-$0.94B | 使本轮可以辩护,但仍略领先于公开证据能证明的水平。 |
ARR 区间是情景假设,不是公司披露指标。
[CV045, CV046, CV047, CV053, CV054, CV055]本轮估值落在基准区间内,但高于公开证据情景的概率加权中心。
区间由情景驱动,并明确受未披露 ARR 和变现假设约束。
[CV053, CV054, CV055, CV056]8.4 建议、投资逻辑破裂点与最终尽调
因此,正确判断是带条件兴趣的观察 / 继续研究,而不是泛泛为「高质量故事」买单。Socket 有足够的品类拉力、产品差异化、客户证据和投资人质量,让 $1 billion 估值变得说得通。但公开证据没有给出决定价格是合理还是偏高的核心指标:ARR、净留存、毛利率、烧钱效率、付费席位转化和优先股堆叠压力。如果管理层能证明当前 ARR 已在数千万美元中段、扩张行为强劲,并且企业 ARPU 能验证溢价模块,这个建议会明显更积极。 反向逻辑不是 Socket 没有需求,而是平台捆绑压缩变现的速度,可能快过单点解决方案扩张的速度。GitHub 明确把 Advanced Security 营销为开发者已在使用的工作流里的原生 AppSec,Copilot 也在扩大对 AI 高占比开发者工作流的控制。如果 Socket 从免费或低价使用转成大型企业合同的能力弱于隐含预期,估值会很快压缩。因此,投资逻辑破裂点可以度量:ARR 低于大约 $40 million,受保护开发者基数上的付费转化率只有低个位数,或有证据显示捆绑的 GitHub 工作流正在赢得基于席位的控制平面。在这些事实解决前,$1 billion 本轮在最佳情况下也只能视为有条件合理;单看公开证据,还略显偏高。[CV057, CV058, CV059, CV060, CV061, CV062]
| 触发项 | 阈值 | 如何传导到投资假设 | 行动含义 |
|---|---|---|---|
| ARR 不足 | 当前 ARR 明显低于 ~$40M | 会打破 $1B 可由上市中档以上倍数支撑的逻辑。 | 重估到悲观区间;除非价格大幅重置,否则放弃。 |
| 付费转化弱 | 受保护开发者变现率为低个位数,或企业 ACV 偏弱 | 显示庞大覆盖面更多是漏斗顶部,而非可变现需求。 | 把本轮从大致合理下调为偏贵。 |
| 打包压力 | GitHub GHAS / Copilot 在核心企业工作流中替代 Socket | 压缩挂载率,并削弱长期席位所有权。 | 下调目标倍数,把平台风险视为打破投资假设。 |
| 毛利 / 烧钱效率弱 | 毛利或烧钱效率明显低于高溢价软件常态 | 把品类故事变成资本强度问题。 | 收紧估值区间,并要求更强融资条款。 |
| 优先权悬顶 | 股权结构或结构化条款在 $1B-$1.2B 结果附近吞掉价值 | 可辩护的企业价值仍可能带来糟糕股权回报。 | 没有清算瀑布清晰度或更优结构,就不要投。 |
触发项是可衡量的经营或结构事件,不是泛泛风险。
[CV048, CV056, CV058, CV059, CV060]| 主题 | 缺失证据 | 重要性 | 负责人或尽调路径 |
|---|---|---|---|
| 当前 ARR / 增长 | 最新 ARR、增长率,以及按产品和套餐拆分的收入桥 | 这是区分合理和偏贵的最大变量。 | CFO 看板、董事会材料或经审计管理账。 |
| 客户质量 | 付费客户数、席位数、ACV 区间,以及按队列拆分的净留存 | 把已披露覆盖面转成真实变现证据。 | 销售 / FP&A 队列切片和前 50 大客户复核。 |
| 单位经济 | 毛利、烧钱、销售效率和云托管成本结构 | 决定高溢价私营倍数能否持续。 | 财务尽调和经营计划复核。 |
| 竞争证据 | 相对 GHAS、GitLab、JFrog 及其他打包替代方案的赢单 / 输单和续约数据 | 检验工作流所有权是持久还是暂时。 | 销售运营分析,加客户背调电话。 |
| 资本结构 | 完全稀释股权表、优先权条款,以及任何老股或要约收购经济性 | 把企业价值转成投资人实际回报数学。 | 法务 + 财务复核股权表、融资文件和 409A 材料。 |
这些问题是把 $1B 当成高信念进入点前必须拿到的最低材料包。
[CV049, CV052, CV060, CV061, CV062]Socket 在市场拉力和产品差异化上得分较高,但披露经济性和估值证据明显弱得多。
评分是 IC 风格的方向性判断,只基于保留的公开证据。
[CV006, CV012, CV049, CV050, CV051, CV058]免责声明
本尽调报告仅基于截至 2026-05-24 的公开信息,不构成投资建议。Socket 是一家私营公司,若干关键财务和合同输入——包括年经常性收入(ARR)、GAAP 收入、毛利率、净留存率(NRR)、烧钱速度、股权结构条款和付费客户数——均未公开披露。公司声称的运营指标、客户名单和产品性能说法,未必能直接映射到付费收入或经审计结果。因此,分析判断和估值区间应视为方向性,而非定论。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Socket describes itself as a developer-first security platform focused on defending software supply chains and open-source dependencies. | 高 | SO002, SO006 |
| CO002 | Socket says it analyzes dependency behavior in real time rather than relying only on known-vulnerability databases after public disclosure. | 高 | SO005, SO006, SO017 |
| CO003 | Socket's official About page says the company was founded in 2021. | 中 | SO002 |
| CO004 | Multiple 2026 funding materials describe Socket as founded in 2020, creating a public-source mismatch with the About page. | 中 | SO006, SO013, SO015, SO017 |
| CO005 | Public company materials and independent funding coverage place Socket in San Francisco, California. | 高 | SO004, SO013, SO017 |
| CO006 | Feross Aboukhadijeh is Socket's founder and CEO. | 高 | SO002, SO006, SO021 |
| CO007 | Feross's public background spans WebTorrent, StandardJS, Node.js governance, and Stanford teaching, giving him unusually strong founder-market fit for open-source supply chain security. | 中 | SO002, SO021, SO022 |
| CO008 | Socket is still hiring across engineering, sales, and customer success, indicating ongoing post-Series-C team expansion. | 中 | SO003, SO005 |
| CO009 | Socket's current product surface includes GitHub integration, a CLI, a VS Code extension, a REST API, a JavaScript SDK, and Socket Firewall. | 中 | SO011, SO023 |
| CO010 | Socket says private source code never leaves the customer's computer or CI environment, and that only dependency lists are sent to its service. | 高 | SO010, SO012 |
| CO011 | Socket's 2026 materials say the company protects more than 27,000 organizations. | 高 | SO001, SO005 |
| CO012 | Socket's 2026 materials say the platform protects 1.5 million repositories and secures more than 11.6 million commits each month. | 高 | SO001, SO005 |
| CO013 | Socket says it blocks more than 10,000 supply-chain attacks each week as of May 2026. | 高 | SO001, SO005 |
| CO014 | In its October 2024 Series B announcement, Socket said it protected more than 7,500 organizations and 300,000 GitHub repositories. | 中 | SO004 |
| CO015 | By April 2025, Socket and acquisition coverage said the company protected more than 8,500 organizations and 750,000+ repositories, scanning every commit in real time and blocking 500+ attacks per week. | 高 | SO007, SO019, SO020 |
| CO016 | Socket said revenue had more than tripled year over year by the time it announced the Coana acquisition. | 中 | SO007 |
| CO017 | By May 2026, Socket said the team had grown to more than 100 people. | 中 | SO005 |
| CO018 | Socket's 2026 funding materials name Anthropic, xAI, Replit, Cursor, Figma, Vercel, Gusto, Mercado Libre, and Cribl as customers. | 高 | SO005, SO006, SO013, SO015, SO017 |
| CO019 | Socket's docs and 2024 customer quotes independently show adoption by Vercel, Replit, Brave, Anthropic, Figma, and MetaMask- or Next.js-adjacent open-source teams. | 中 | SO004, SO011 |
| CO020 | Socket announced a $40 million Series B on 2024-10-22 led by Abstract Ventures. | 高 | SO004, SO007 |
| CO021 | Socket said the Series B brought cumulative funding to $65 million. | 中 | SO004 |
| CO022 | Socket announced a $60 million Series C at a $1 billion valuation on 2026-05-20. | 高 | SO005, SO006, SO013, SO014 |
| CO023 | Thrive Capital led Socket's Series C, with participation from Andreessen Horowitz, Abstract Ventures, and Capital One Ventures. | 高 | SO005, SO006, SO013, SO014, SO015, SO016 |
| CO024 | Socket said the Series C brought total funding to $125 million. | 高 | SO005, SO006, SO016, SO017 |
| CO025 | Socket said Series C proceeds would fund Firewall expansion, Certified Patches, broader ecosystem coverage, enterprise growth, and new product launches. | 高 | SO005, SO006 |
| CO026 | Socket frames AI-generated code as a demand accelerator because it increases the volume of third-party dependencies reaching production. | 高 | SO005, SO006, SO015, SO017 |
| CO027 | Socket announced the acquisition of Coana on 2025-04-25 to add reachability analysis and static/control-flow analysis to the platform. | 高 | SO007, SO020 |
| CO028 | Socket and Coana say the reachability engine can eliminate up to 80% of false positives and improve remediation speed by up to 10x. | 高 | SO007, SO018, SO019, SO020 |
| CO029 | The entire Coana team joined Socket as part of the acquisition. | 高 | SO007, SO018, SO020 |
| CO030 | By 2026, Socket said Secure Annex was its second acquisition in 12 months and that it extended coverage from package managers into browser extensions, code editor extensions, MCP servers, and AI tools. | 高 | SO008, SO005 |
| CO031 | Socket Firewall Free launched on 2025-09-30 as a free install-time protection tool for JavaScript/TypeScript, Python, and Rust package managers. | 高 | SO009, SO024, SO026 |
| CO032 | Socket Firewall blocks malicious packages by acting as a proxy between package managers and registries, checking packages before download and applying policy to direct and transitive dependencies. | 高 | SO009, SO012, SO024, SO026 |
| CO033 | The free Firewall product warns on AI-detected malware but does not auto-block unconfirmed AI-only flags, while enterprise adds configurable policy, custom registries, allow-lists, and broader ecosystem coverage. | 高 | SO009, SO012, SO024, SO026 |
| CO034 | Socket's docs still position GitHub integration as the easiest entry point, with CLI and other interfaces as alternate workflows rather than separate businesses. | 中 | SO011, SO023 |
| CO035 | Socket has publicly acknowledged that AI-assisted malware detection can create false positives, which is why the free firewall defaults to warning rather than blocking AI-only signals. | 高 | SO009, SO026 |
| CO036 | A public March 2026 GitHub issue reported Socket flagging harmless RFC 2606 example-domain strings as a supply-chain risk, showing that at least some false-positive complaints reach end users. | 中 | SO025 |
| CO037 | Independent coverage frames Socket as competing against Snyk, Checkmarx, Sonatype, and GitHub, so category leadership is still an execution claim rather than a settled market fact. | 中 | SO017 |
| CO038 | Socket's public narrative remains highly founder-centric, making Feross Aboukhadijeh a meaningful key-person dependency for product vision, customer credibility, and recruiting. | 中 | SO002, SO004, SO005, SO006 |
| CO039 | Socket says the core product remains free for open-source projects while paid plans monetize enterprise needs such as invoicing, volume discounts, and premium support. | 中 | SO010 |
| CO040 | Socket Firewall Free collects anonymous telemetry, while enterprise deployments let organizations configure telemetry controls. | 高 | SO009, SO012, SO026 |
| CO041 | Socket said it identified the malicious Axios dependency within six minutes and onboarded more than 2,000 organizations within 24 hours of the incident. | 高 | SO005, SO006, SO015, SO017 |
| CO042 | Socket's current platform breadth spans install-time blocking, dependency analysis, reachability triage, and GitHub or CLI workflows rather than a single scanner product. | 中 | SO011, SO012, SO023, SO024 |
| CO043 | Socket's About page emphasizes a backer roster that includes a16z, Abstract Ventures, Elad Gil, Bret Taylor, Patrick Collison, John Collison, Ryan Dahl, and other security or open-source operators. | 中 | SO002, SO004 |
| CO044 | Reviewed public materials do not disclose Socket's revenue or ARR, board composition, debt, or secondary-liquidity details with enough precision for a full capitalization model. | 中 | SO002, SO010, SO013, SO017 |
| CM001 | Open source dependencies are pervasive enough that dependency risk is a structural software problem, not a niche corner case. | 中 | SM001, SM015, SM032 |
| CM002 | Socket positions itself as a developer-first platform for vulnerable and malicious dependencies rather than as a full application security suite. | 中 | SM001, SM002 |
| CM003 | The direct market includes dependency admission control, pull-request gating, malicious package detection, SBOM-aware inventory, and advisory triage inside software delivery workflows. | 中 | SM002, SM004, SM023, SM030 |
| CM004 | Status-quo alternatives include built-in repo-host features, open vulnerability data, CVE scanners, SBOM platforms, and automated dependency update bots. | 中 | SM021, SM026, SM027, SM029, SM030, SM031 |
| CM005 | Socket highlights non-CVE supply chain signals such as typosquats, install scripts, obfuscation, shell access, environment-variable access, and network activity. | 中 | SM004 |
| CM006 | Dependabot is available for all GitHub repositories and automates both version updates and security updates through pull requests. | 中 | SM021 |
| CM007 | npm audit gives JavaScript teams a built-in package security audit without buying a separate commercial tool. | 中 | SM026 |
| CM008 | OSV provides open vulnerability data plus scanner workflows for lockfiles, SBOMs, images, and CI/CD usage. | 中 | SM027 |
| CM009 | OWASP Dependency-Check is a software composition analysis tool that maps dependencies to publicly disclosed vulnerabilities. | 中 | SM029 |
| CM010 | Dependency-Track consumes and analyzes SBOMs and aggregates multiple vulnerability data sources, showing that inventory and policy workflows sit inside the direct category. | 中 | SM030 |
| CM011 | Broader AppSec platforms such as Black Duck Polaris package SAST, SCA, and DAST together, so much application security spend is adjacent to Socket rather than directly comparable. | 中 | SM011, SM018, SM019 |
| CM012 | Verified Market Reports sizes software supply chain security at USD 1.2 billion in 2025 growing to USD 4.5 billion by 2034 at a 16.5% CAGR. | 中 | SM017 |
| CM013 | Mordor sizes the broader application security market at USD 14.83 billion in 2026 growing to USD 28.11 billion by 2031. | 中 | SM019 |
| CM014 | Fortune sizes the broader application security market at USD 14.86 billion in 2026 growing to USD 43.28 billion by 2034. | 中 | SM020 |
| CM015 | Mordor's SCA page claims a USD 430.12 billion market in 2026, which is dramatically larger than adjacent AppSec estimates. | 低 | SM015 |
| CM016 | Public market estimates therefore span a direct low-single-digit-billions lens, a broader mid-teens-billions AppSec adjacency, and at least one clearly over-broad SCA estimate. | 中 | SM015, SM017, SM019, SM020 |
| CM017 | A Socket-relevant SAM is best framed as recurring developer or committer coverage for dependency control workflows rather than as all application security spend. | 中 | SM003, SM022, SM025 |
| CM018 | Socket measures a billable developer as someone who committed to a scanned repository in the past 90 days. | 中 | SM003 |
| CM019 | GitHub Code Security is priced at USD 30 per active committer per month and Secret Protection at USD 19 per active committer per month. | 中 | SM022 |
| CM020 | GitLab packages richer dependency security in its Ultimate enterprise tier aimed at advanced security and compliance use cases. | 中 | SM023, SM024 |
| CM021 | Snyk segments plans from free individual use to team, Ignite, and enterprise tiers and exposes SCA-related usage quotas and SBOM support. | 中 | SM025 |
| CM022 | Renovate offers multi-platform automated dependency update pull requests as open source or via Mend-hosted service, making it a low-cost alternative for update automation. | 中 | SM031 |
| CM023 | Known-vulnerability scanning and SBOM analysis are partially commoditized because OSV, npm audit, Dependency-Check, and Dependency-Track are available at low or no direct software cost. | 中 | SM026, SM027, SM029, SM030 |
| CM024 | GitHub Advanced Security explicitly argues that dependency security works inside native GitHub workflows rather than as a third-party add-on, which is a major distribution advantage. | 中 | SM022 |
| CM025 | GitLab recommends SBOM-based dependency scanning for new projects and continuously rescans SBOM components when advisories change. | 中 | SM023 |
| CM026 | GitLab is experimenting with analyzing dependencies for behaviors to surface suspicious or malicious activity beyond known CVEs. | 中 | SM023 |
| CM027 | The category frontier is moving from CVE-only scanning toward behavior-aware, context-aware, and continuously rescanned dependency risk. | 中 | SM004, SM023, SM027 |
| CM028 | EO 14028 tasked NIST with initiatives related to the security and integrity of the software supply chain. | 中 | SM008 |
| CM029 | NIST SSDF says secure development practices reduce released vulnerabilities and can be used by purchasers and consumers in acquisition processes. | 中 | SM006 |
| CM030 | CISA describes the SBOM as a key building block in software security and software supply chain risk management. | 中 | SM007 |
| CM031 | The EU Cyber Resilience Act imposes lifecycle cybersecurity requirements and starts reporting obligations on 11 September 2026. | 中 | SM033 |
| CM032 | Together, EO 14028, SSDF, SBOM policy, and the CRA make software supply chain evidence increasingly procurement-relevant rather than optional hygiene. | 中 | SM006, SM007, SM008, SM033 |
| CM033 | The XZ incident showed that upstream xz tarballs and liblzma could be backdoored in ways that affected ssh server compromise paths. | 中 | SM009 |
| CM034 | Apache's Log4j security page still documents upgrade guidance around CVE-2021-44228 and later fixes, illustrating the long remediation tail of transitive dependency incidents. | 中 | SM010 |
| CM035 | High-profile incidents such as XZ and Log4Shell keep software supply chain security on executive and procurement agendas. | 中 | SM008, SM009, SM010 |
| CM036 | Sonatype says repository abuse accounted for 55.9% of logged malicious packages and secrets exfiltration appeared in 3.9%, showing attacker focus on developer and CI contexts. | 中 | SM012 |
| CM037 | Sonatype also reports droppers or loaders, backdoors, and obfuscated code in malicious packages, indicating chained attacks rather than one-off payloads. | 中 | SM012 |
| CM038 | Veracode says npm represented 65.9% of the malicious packages it saw and recorded 42,313 malicious-URL packages, 89,373 suspicious install-code packages, 555,258 obfuscated packages, and 4,708 typosquats in the period. | 中 | SM014 |
| CM039 | Veracode says malicious URLs rose 179.2% and typosquats 104.3%, suggesting attackers are leaning harder into developer deception and package admission mistakes. | 中 | SM014 |
| CM040 | Mordor attributes SCA demand to SBOM and compliance mandates, supply-chain attacks, shift-left DevSecOps budgets, and AI-generated transitive dependencies. | 中 | SM015 |
| CM041 | Mordor says large enterprises held 72.9% of 2025 SCA revenue while SMEs were fastest-growing, and IT and telecom led current demand while healthcare and life sciences grew fastest. | 中 | SM015 |
| CM042 | Market adoption is constrained by false-positive fatigue, talent shortages, total cost of ownership, and tool sprawl. | 中 | SM015, SM019, SM020 |
| CM043 | GitHub, GitLab, Snyk, and broader AppSec platforms compress the direct market by bundling dependency security inside existing platforms and contracts. | 中 | SM011, SM022, SM023, SM024, SM025 |
| CM044 | Built-in and open-source substitutes commoditize known-vulnerability scanning and inventory, so premium vendors must win on precision, malicious-package detection, workflow fit, or compliance depth. | 中 | SM021, SM026, SM027, SM029, SM030, SM031, SM004 |
| CM045 | The highest-fit premium segment is organizations that start with free or bundled tools and upgrade when central security needs policy, reduced noise, malicious-package detection, or compliance evidence. | 中 | SM003, SM004, SM022, SM023, SM024, SM025 |
| CP001 | Socket Firewall stops supply chain attacks at install time by intercepting package downloads and enforcing policy on developer machines, CI pipelines, and networks. | 中 | SP002 |
| CP002 | Socket publishes four pricing tiers: Free '$0' per developer per month, Team '$25', Business '$50', and custom Enterprise pricing. | 中 | SP001 |
| CP003 | Socket's public paid tiers extend beyond basic alerting because Team adds precomputed reachability and Slack alerts while Business adds SBOM import and export, SSO or SAML, webhook automation, GitHub Actions scanning, and AI model scanning. | 中 | SP001 |
| CP004 | Socket says full application reachability scans both app source and dependency code, can mark around 80% of vulnerabilities irrelevant, can exceed 90% noise reduction in some ecosystems, and is compute-intensive enough that customers often enable it selectively. | 中 | SP001, SP004 |
| CP005 | Socket's clearest public differentiation is behavior-based malicious-package blocking before download or execution rather than a broad code-to-cloud platform story. | 中 | SP002, SP023 |
| CP006 | Snyk sells a single platform across open source, code, container, IaC, API or web, and AI security workflows. | 中 | SP005, SP006 |
| CP007 | Snyk Open Source emphasizes developer-first integration across IDEs, repos, CI or CD, and live environments, with prioritization that factors reachability, exploit maturity, and EPSS or CVSS. | 中 | SP006 |
| CP008 | Snyk prices by contributing developer and keeps public Free, Team, Ignite, and Enterprise plan tiers. | 中 | SP005 |
| CP009 | Mend AppSec markets a broader platform than Socket by combining code, dependency, container, AI-code, and automated dependency-update capabilities under one product family. | 中 | SP007 |
| CP010 | Mend explicitly prices per contributing developer and says pricing does not increase with code size, number of scans, or number of applications. | 中 | SP007 |
| CP011 | Endor Labs AURI markets full-stack reachability across first-party code, transitive dependencies, and container images and claims up to 95% noise reduction. | 中 | SP008 |
| CP012 | Endor Labs says its MCP, Skills, and CLI are free for individual developers while organization-wide policies, governance, and integrations sit in the enterprise platform. | 中 | SP008 |
| CP013 | Endor Labs' competitive page argues that Socket's package signals can feel opaque and that Endor offers a more transparent and customizable policy engine. | 中 | SP009 |
| CP014 | JFrog Xray is an enterprise SCA tool that continuously scans repositories, build packages, and container images and includes license compliance, SBOMs, and malicious-package detection. | 中 | SP010 |
| CP015 | Xray is a core component of JFrog Platform subscriptions and is included with Pro X, Enterprise X, or Enterprise+ rather than sold as a freemium developer add-on. | 中 | SP010, SP011 |
| CP016 | FOSSA's public pricing centers on compliance operations, with exported SBOMs in the free tier and snippet-scanning plus binary-scanning add-ons at enterprise scope. | 中 | SP012 |
| CP017 | FOSSA publishes Free, '$20 per project per month' Business, and custom Enterprise plans. | 中 | SP012 |
| CP018 | GitHub Advanced Security sells native GitHub Secret Protection for '$19' per active committer per month and GitHub Code Security for '$30' per active committer per month. | 中 | SP013 |
| CP019 | GitHub frames GHAS as built-in native AppSec inside GitHub workflows instead of a separate third-party toolchain. | 中 | SP013 |
| CP020 | Dependabot alerts notify repository owners about known vulnerable dependencies on the default branch, but GitHub documents that alerts cannot catch every issue and only fire from reviewed advisories in supported ecosystems. | 中 | SP014 |
| CP021 | Apiiro competes as an ASPM and software-supply-chain platform built around a risk graph, contextual SCA, secure-by-design controls, and extended SBOM or XBOM generation. | 中 | SP015 |
| CP022 | Chainguard competes from the hardened-image and library layer with contractual CVE remediation SLAs and catalog pricing that starts at '$19K' for a team of 10. | 中 | SP016 |
| CP023 | Aikido markets one platform across SCA, SAST, IaC, DAST, container scanning, secrets, cloud posture, runtime protection, and dependency malware detection. | 中 | SP017, SP018 |
| CP024 | Aikido pairs its consolidation pitch with public pricing, on-prem deployment options, and explicit migration messaging against tools such as Snyk. | 中 | SP017, SP018 |
| CP025 | OX Security markets a single code-to-cloud platform priced per developer and spanning SAST, SCA, SBOM, Git posture, CI or CD security, runtime, attack-path analysis, and pentesting. | 中 | SP019, SP020 |
| CP026 | Upwind is an adjacent substitute rather than a pure Socket clone because it bundles SCA or SBOM, application security, posture, API security, and runtime protection into a runtime-first cloud and AI platform. | 中 | SP021 |
| CP027 | Pixee's May 2026 market review argues that SCA detection is increasingly commoditized and that the bottleneck has shifted to triage, exploitability context, and remediation. | 中 | SP022 |
| CP028 | The same review maps Snyk to developer-first breadth, Mend to enterprise consolidation, Endor Labs to deep reachability, FOSSA to legal workflows, and Dependabot to free dependency freshness. | 中 | SP022 |
| CP029 | AppSec Santa's 2026 alternatives review describes Socket as especially strong for npm and JavaScript supply-chain attacks but narrower for polyglot estates and platform-consolidation buyers. | 中 | SP023 |
| CP030 | SourceForge's comparison page clusters Endor Labs, Aikido, and Chainguard around the same buyer journey as Socket, showing that buyers comparison-shop across direct SCA and broader supply-chain platforms. | 中 | SP024 |
| CP031 | PeerSpot's 2026 comparison gives Snyk higher AppSec-tools mindshare than GitHub Advanced Security and says Snyk wins on breadth and integrations while GHAS wins on GitHub-native integration. | 中 | SP025 |
| CP032 | Socket's direct competitive set spans four classes: specialist SCA or AppSec vendors, GitHub-native substitutes, compliance or artifact-centric incumbents, and broader code-to-cloud platforms. | 中 | SP022, SP023, SP024 |
| CP033 | Socket's moat is strongest when buyers explicitly value behavior-based malicious-package detection and install-time blocking before code ever runs. | 中 | SP002, SP023 |
| CP034 | Socket is more price-transparent than most enterprise quote-led rivals, but its paid seats still stack on top of free or native GitHub baselines rather than replacing them by default. | 中 | SP001, SP013, SP014 |
| CP035 | GitHub-native dependency monitoring is the clearest low-cost substitute for Socket in GitHub-centered teams because it already lives in the repository workflow and covers known-vulnerability freshness. | 中 | SP013, SP014, SP022 |
| CP036 | Endor Labs and Snyk are the clearest direct pressure on Socket's noise-reduction story because both market reachability, exploitability context, and fix workflows rather than only package reputation signals. | 中 | SP006, SP008 |
| CP037 | Aikido, OX Security, Apiiro, and Upwind pressure Socket from the consolidation side by combining dependency security with code, cloud, API, or runtime coverage under one contract. | 中 | SP015, SP018, SP019, SP021 |
| CP038 | FOSSA and JFrog Xray pressure Socket in compliance-heavy and artifact-centric environments where SBOM, license, binary, and registry workflows matter more than npm-first malware analysis. | 中 | SP010, SP012 |
| CP039 | Chainguard is more substitute than direct peer because it shifts the control point to trusted images and libraries with contractual remediation SLAs, which matters most in container-heavy regulated environments. | 中 | SP016 |
| CP040 | Competitive risk is highest if buyers conclude that bundled or broader platforms deliver enough supply-chain coverage without adding another paid specialist. | 中 | SP022, SP023, SP025 |
| CP041 | Socket's switching costs are moderate rather than extreme because the product can layer into existing repo workflows, which makes multi-homing possible even when Socket wins the specialist slot. | 中 | SP001, SP013, SP014 |
| CP042 | GitHub has the strongest workflow distribution advantage in this category because GHAS and Dependabot surface directly inside the repo and security tab many developers already use daily. | 中 | SP013, SP014, SP025 |
| CP043 | Socket's paid scope is broader than a pure alerting scanner because it now includes reachability, SBOM support, GitHub Actions scanning, AI model scanning, and Firewall, but it still stops short of the code-to-cloud breadth claimed by Aikido, OX Security, or Upwind. | 中 | SP001, SP002, SP019, SP021 |
| CP044 | OX Security and Upwind both explicitly market multiple-tool replacement and code-to-cloud or runtime visibility, raising the proof burden on any specialist tool seeking a separate budget line. | 中 | SP019, SP021 |
| CP045 | FOSSA and Chainguard both publish adjacent-category price anchors, giving buyers transparent alternatives to opaque enterprise quotes elsewhere in the market. | 中 | SP012, SP016 |
| CP046 | GitHub Advanced Security pricing creates a public per-active-committer price anchor inside the same workflow many Socket prospects already use. | 中 | SP013, SP025 |
| CP047 | For lower-complexity teams, the practical status-quo substitute is Dependabot-style alerting plus internal package-governance process rather than a standalone specialist purchase. | 中 | SP014, SP022 |
| CI001 | Socket said on 2026-05-20 that it raised $60 million in Series C funding at a $1 billion valuation. | 高 | SI001, SI002, SI008, SI009, SI010 |
| CI002 | The May 2026 Series C was led by Thrive Capital with participation from Andreessen Horowitz, Abstract Ventures, and Capital One Ventures. | 高 | SI001, SI002, SI008, SI009, SI010 |
| CI003 | Socket’s total disclosed funding reached $125 million after the Series C. | 高 | SI001, SI002, SI010, SI015 |
| CI004 | Socket’s Series C blog says the company grew from 7,500 organizations at Series B close to more than 27,000 organizations by May 2026. | 中 | SI002 |
| CI005 | Official May 2026 materials say Socket protects 1.5 million repositories and secures more than 11.6 million commits every month. | 高 | SI002, SI006 |
| CI006 | Socket’s Series C blog says the team has grown to more than 100 people. | 中 | SI002 |
| CI007 | Socket’s public Series C materials name Anthropic, xAI, Replit, Cursor, Figma, Vercel, Gusto, Mercado Libre, and Cribl as customers alongside Fortune 100 companies. | 高 | SI001, SI002, SI008, SI009 |
| CI008 | Socket’s public list pricing is Free at $0, Team at $25 per developer per month, Business at $50 per developer per month, and Enterprise custom. | 中 | SI003 |
| CI009 | Socket says annual billing saves up to 20 percent and enterprise plans can receive volume-based discounts and manual invoicing. | 中 | SI003 |
| CI010 | Socket defines a billable developer as someone who made a commit to an organization repository scanned by Socket in the past 90 days. | 中 | SI003 |
| CI011 | Socket says open-source projects remain free and early-stage startups can request special pricing. | 中 | SI003 |
| CI012 | The pricing page gives Free 1,000 scans per month, Team 5,000 scans per month, and Business unlimited scans and API quota. | 中 | SI003 |
| CI013 | Business and Enterprise pricing include compliance integrations, SBOM workflows, SSO/SAML, audit logs, and higher-touch support features. | 中 | SI003 |
| CI014 | Socket’s October 2024 Series B raised $40 million and was led by Abstract Ventures with participation from Elad Gil and Andreessen Horowitz. | 高 | SI011, SI012, SI013, SI014 |
| CI015 | Public Series B coverage says that round took Socket’s total funding to $65 million. | 高 | SI012, SI013, SI014 |
| CI016 | Tracxn says Socket’s first funding round occurred in May 2022 and that the company had completed four rounds by May 2026. | 中 | SI015 |
| CI017 | Most funding and market-data sources in the retained set place Socket’s founding in 2020. | 高 | SI001, SI002, SI008, SI009, SI011, SI015 |
| CI018 | Socket’s About page instead says the company was founded in 2021, creating an inconsistency in public profile data. | 低 | SI004 |
| CI019 | Socket’s monetization is subscription SaaS priced per developer with annual-prepay, invoice, and marketplace procurement options rather than one-time licensing. | 中 | SI003 |
| CI020 | The public price list reveals contract architecture but not realized ACV, discounting, or renewal quality. | 中 | SI003, SI001, SI002 |
| CI021 | Socket’s homepage and Series C blog both report more than 27,000 organizations protected. | 高 | SI002, SI006 |
| CI022 | Socket’s homepage and Series C blog both report more than 10,000 attacks blocked every week. | 高 | SI002, SI006 |
| CI023 | Socket’s homepage says the company protects 1.5 million code repositories and secures 11.6 million or more commits every month. | 高 | SI002, SI006 |
| CI024 | Pricing and packaging imply a self-serve land motion with enterprise upsell into compliance, reachability, and support-heavy contracts. | 中 | SI003, SI007 |
| CI025 | Socket’s careers page emphasizes competitive salary benchmarking, stock options, insurance, remote work, and quarterly offsites, implying continued people investment. | 中 | SI005 |
| CI026 | Socket’s careers page links to an Ashby jobs board, showing public recruiting infrastructure remained live in May 2026. | 中 | SI005, SI026 |
| CI027 | ZoomInfo models Socket at about $18.1 million of revenue and 51-200 employees, but those are third-party estimates rather than company disclosures. | 低 | SI016 |
| CI028 | Tracxn shows Socket as a Series C company with $125 million raised but hides key valuation and operating details behind gated fields. | 低 | SI015 |
| CI029 | Retained public sources do not disclose Socket’s ARR, GAAP revenue, gross margin, NRR, cash balance, or runway months. | 中 | SI001, SI002, SI003, SI005, SI006 |
| CI030 | No retained public source disclosed venture debt, project finance, or secondary share-sale terms for Socket. | 中 | SI001, SI002, SI010, SI011, SI012, SI013, SI015 |
| CI031 | Socket framed the Series C as funding to scale the platform, expand enterprise adoption, and secure the software supply chain as AI accelerates development. | 高 | SI001, SI002, SI008, SI009 |
| CI032 | Socket announced the Coana acquisition on 2025-04-23, and official plus independent coverage agree that the purchase price was undisclosed. | 高 | SI007, SI017, SI020, SI021 |
| CI033 | Socket says Coana’s reachability engine can cut false positives by up to 80 percent and improve remediation speed by up to 10x. | 中 | SI007 |
| CI034 | Tech Funding News estimated Coana’s purchase price at $50 million to $100 million, but the range is analyst speculation rather than disclosed consideration. | 低 | SI019 |
| CI035 | Business Partner Magazine and Tech Funding News reported roughly 300 percent year-over-year revenue growth around the Coana acquisition, but the claim is not corroborated in Socket’s official Series C disclosures. | 低 | SI018, SI019, SI001, SI002 |
| CI036 | Forbes presented the Coana deal as the next phase after Socket’s 2024 Series B, supporting a narrative of product-led M&A rather than distressed consolidation. | 中 | SI017, SI011 |
| CI037 | INCUBA says Coana was founded in 2022, backed by Sequoia and others, and exited to Socket in one of the largest exits in the incubator’s environment. | 中 | SI021 |
| CI038 | The Coana rationale centers on reducing alert fatigue and false positives, so the financial upside is more likely retention and upsell than immediately disclosed revenue contribution. | 中 | SI007, SI017, SI019, SI020 |
| CI039 | Scamadviser demonstrates that generic website-trust heuristics are weak diligence inputs because it reports a trust score of zero while also saying socket.dev is likely legitimate. | 低 | SI025 |
| CI040 | Disclosed external capital nearly doubled from $65 million after Series B to $125 million after Series C. | 高 | SI012, SI013, SI014, SI002, SI010 |
| CI041 | A $1 billion valuation on still-private revenue, margin, and cash metrics means public underwriting rests more on growth narrative and customer quality than on auditable unit economics. | 中 | SI001, SI002, SI010, SI016 |
| CI042 | Socket’s price page offers separately purchasable products and enterprise-only features, implying a multi-product expansion path beyond base dependency scanning. | 中 | SI003 |
| CI043 | Enterprise support features such as private Slack, account management, migration help, audit logs, and SCIM imply meaningful service-delivery costs for large accounts. | 中 | SI003 |
| CI044 | Public pricing and product delivery point to a capital-light software model rather than hardware or inventory-heavy economics. | 中 | SI003, SI004, SI006 |
| CI045 | Socket uses its funding and customer credentials as commercial proof points on public pages, which may help sales efficiency but does not substitute for disclosed realized pricing or retention. | 中 | SI003, SI006 |
| CI046 | Cooley advised Socket on both the 2024 Series B and 2026 Series C, which is consistent with standard venture-equity financing rather than unusual structured capital. | 中 | SI010, SI012 |
| CI047 | Socket’s Enterprise plan can be purchased through GCP Marketplace, adding another procurement path for larger customers. | 中 | SI003 |
| CI048 | The financial logic of the Coana deal is not just feature breadth; it is lowering alert noise so customers can focus on exploitable issues, which should improve product ROI if the claim holds in practice. | 中 | SI007, SI017, SI019, SI020 |
| CI049 | Even after the Coana acquisition, public sources do not disclose purchase consideration, integration cost, or synergy timing, so capital-allocation quality is only partially underwritten. | 中 | SI007, SI017, SI019 |
| CI050 | Public evidence is sufficient to map pricing, financing history, and traction, but not sufficient to fully underwrite realized revenue quality, margin, or cash resilience at the current valuation. | 中 | SI001, SI002, SI003, SI016 |
| CE001 | Socket positions the product as blocking malicious packages before they reach code rather than only ranking dependency risk after the fact. | 中 | SE001, SE003 |
| CE002 | The public product surface spans a GitHub app, CLI, VS Code extension, Firewall, REST API, and SDKs rather than a single scanning interface. | 中 | SE002 |
| CE003 | Socket for GitHub analyzes newly added or updated dependencies in pull requests and posts review output before code is merged. | 中 | SE004, SE005 |
| CE004 | Official GitHub marketing presents the GitHub app as the easiest entry point and a two-click installation flow. | 中 | SE002, SE004 |
| CE005 | The CLI is the lower-level workflow for teams that want more control or do not rely on GitHub. | 中 | SE002, SE016 |
| CE006 | The VS Code extension lets developers scan package manifest files inside the editor and receive immediate security feedback. | 中 | SE002, SE006 |
| CE007 | The VS Code docs say some extension analysis depends on the Socket API and an internet connection, so the editor workflow is not a full offline replacement for all checks. | 中 | SE006 |
| CE008 | Socket Firewall intercepts direct and transitive dependency installs at install time and can block malicious packages before execution. | 中 | SE002, SE007 |
| CE009 | Firewall is designed for developer machines, CI pipelines, and network choke points with centralized policy and telemetry. | 中 | SE007 |
| CE010 | Socket’s FAQ publicly claims support across JavaScript, Python, Java, Ruby, .NET, Go, Rust, Scala, and Kotlin, with additional ecosystems planned. | 中 | SE003 |
| CE011 | Firewall marketing specifically calls out JavaScript, Python, Rust, and enterprise support for Maven, Ruby, NuGet, and beyond. | 中 | SE007 |
| CE012 | The GitHub feature page currently spotlights JavaScript, Python, and Go dependencies in the PR workflow. | 中 | SE004 |
| CE013 | The technical core combines package behavior analysis, package metadata analysis, and maintainer-behavior analysis. | 中 | SE003 |
| CE014 | Public docs say Socket inspects behaviors such as network access, filesystem access, shell execution, environment-variable reads, install scripts, obfuscation, and telemetry. | 中 | SE003, SE029 |
| CE015 | Socket’s FAQ says it looks for 70-plus signals, while its GitHub Marketplace page lists 70 detections across six categories. | 中 | SE003, SE005 |
| CE016 | Socket explicitly positions its design against CVE-only tooling by saying malicious behavior can be identified before public vulnerability disclosure. | 中 | SE003, SE021 |
| CE017 | AppSec Santa characterizes Socket as supply-chain-focused SCA that is distinct from Dependabot- or Snyk-style CVE-first approaches. | 中 | SE022 |
| CE018 | The GitHub Marketplace listing shows Socket categories spanning supply chain risk, vulnerability, quality, maintenance, and license issues rather than vulnerability alerts alone. | 中 | SE005 |
| CE019 | Reachability is now a first-class product surface with both a dedicated feature page and dedicated technical documentation. | 中 | SE008, SE009 |
| CE020 | Socket markets three reachability tiers: full application reachability, precomputed reachability, and dependency reachability. | 中 | SE008 |
| CE021 | The reachability surface is marketed as cutting up to 90%, 80%, and 35% of irrelevant or unreachable CVE noise across the three tiers, respectively. | 中 | SE008, SE011 |
| CE022 | Full application reachability requires a CLI or GitHub Action setup, unlike the lower-friction precomputed tier that works across existing integrations. | 中 | SE008, SE009 |
| CE023 | The full-application docs say analysis cost scales with language type, program size, dependency graph size, and the number of CVEs under consideration. | 中 | SE009 |
| CE024 | The full-application docs enumerate language-specific requirements such as Python 3.11+, .NET 6+, matching Go versions, and lockfile or SBOM prerequisites for some Java/Gradle flows. | 中 | SE009 |
| CE025 | Socket cites the Coana acquisition as the mechanism that brought advanced static and control-flow reachability analysis into the platform. | 中 | SE011, SE012, SE023, SE024 |
| CE026 | Socket says the Coana integration adds precomputed reachability that can suppress unused transitive vulnerability alerts without source-code upload in the demo flow. | 中 | SE011, SE012 |
| CE027 | Socket says function-level reachability can run on the user’s machine or CI runner and can even operate fully offline on an air-gapped network. | 中 | SE011, SE012 |
| CE028 | The product-news feed shows a 2025-2026 release cadence that includes OpenVSX scanning, Ruby reachability beta, Immutable Scans, PHP and Composer support, Jira, and Data Exports. | 中 | SE013 |
| CE029 | Socket’s research feed shows active detection work across npm, Go, NuGet, RubyGems, Packagist, PyPI, extension ecosystems, and CI-oriented attack paths in 2026. | 中 | SE014 |
| CE030 | GitHub’s organization API shows Socket maintained 46 public repositories and 712 followers as of 2026-05-19. | 中 | SE015 |
| CE031 | The public socket-cli repository was updated on 2026-05-23 and had 271 stars at fetch time. | 中 | SE016 |
| CE032 | The public socket-vscode repository was updated on 2026-05-21 and had 21 stars at fetch time. | 中 | SE017 |
| CE033 | The public socket-sdk-js repository was updated on 2026-05-23 and had 50 stars at fetch time. | 中 | SE018 |
| CE034 | The public socket-sdk-python repository was updated on 2026-05-22 and had 12 stars at fetch time. | 中 | SE019 |
| CE035 | Socket’s homepage and 2026 Series C materials claim 27,000-plus organizations, 1.5 million repositories, 11.6 million commits per month, and 10,000-plus blocked attacks per week. | 中 | SE001, SE020, SE021, SE025, SE026, SE027 |
| CE036 | Series C materials describe Firewall, reachability, and Certified Patches as flagship product-expansion areas. | 中 | SE020 |
| CE037 | The 2026 Series C post says Socket is extending protection from package managers into browser extensions, code editor extensions, MCP servers, and AI tools. | 中 | SE020 |
| CE038 | Official pricing and FAQ language say private source code stays on the developer machine or CI environment and that Socket primarily receives manifests and dependency lists. | 中 | SE003, SE010 |
| CE039 | Socket’s FAQ says the service does not process PII or analyze proprietary customer source code. | 中 | SE003 |
| CE040 | Independent reviewers describe paid plans and free-tier limits as practical adoption constraints for larger organizations. | 中 | SE028, SE029 |
| CE041 | Ry Walker Research says the strongest public fit today is still primarily JavaScript, Python, and Go and warns that behavioral analysis can create false positives. | 中 | SE028 |
| CE042 | Startupik says coverage outside the core JavaScript workflow is still evolving and that noisy results can appear in dynamic or experimental repositories if policies are not tuned. | 中 | SE029 |
| CE043 | AppSec Santa says teams may still pair Socket with traditional SCA or broader policy and compliance tooling instead of treating it as a one-product replacement. | 中 | SE022 |
| CE044 | Taken together, the product behaves more like a developer-workflow security platform than a pure vulnerability scanner because it combines PR checks, editor feedback, install-time enforcement, API or SDK access, and reachability-guided triage. | 中 | SE002, SE004, SE006, SE007, SE008, SE009 |
| CE045 | The biggest remaining product-tech diligence gaps are public evidence on cross-language feature parity and procurement-grade assurance depth, not a lack of outward product surface or release velocity. | 中 | SE003, SE010, SE028, SE029 |
| CU001 | As of May 2026, Socket says it protects more than 27,000 organizations, 1.5 million repositories, and 11.6 million commits per month. | 高 | SU003, SU020 |
| CU002 | Socket's May 2026 official materials list Anthropic, xAI, Replit, Cursor, Vercel, Figma, Gusto, Mercado Libre, and Cribl among its customers, alongside unnamed Fortune 100 companies in finance and global media. | 高 | SU002, SU003 |
| CU003 | A Thrive Capital partner said Cursor, OpenAI, and Anthropic independently described Socket as the most important security tool they had adopted in response to AI-driven development. | 中 | SU003 |
| CU004 | The reviewed public customer proof clusters around AI-native, cloud, developer-platform, identity, compliance, and security-conscious software organizations rather than a broad offline enterprise base. | 中 | SU001, SU003, SU004, SU005, SU006, SU009, SU010, SU012, SU013, SU015 |
| CU005 | Across named case studies, the buyer is usually a CISO, security engineering leader, or platform-security manager, with the security budget owner sponsoring rollout. | 中 | SU004, SU006, SU010, SU012, SU015 |
| CU006 | The day-to-day users are developers and platform engineers who receive dependency feedback inline in pull requests or GitHub checks. | 高 | SU005, SU008, SU011, SU023 |
| CU007 | Socket's most visible initial deployment motion is a low-friction GitHub App or GitHub-check rollout rather than a heavyweight standalone security-console rollout. | 高 | SU005, SU010, SU023 |
| CU008 | Anthropic embedded Socket's API into its internal dependency approval pipeline so packages meeting thresholds are auto-approved and others are escalated for manual review. | 中 | SU004 |
| CU009 | Anthropic says Socket cut hands-on dependency-review effort by 95% and saves security engineers more than five hours per week. | 中 | SU004 |
| CU010 | Replit describes Socket as a GitHub-check workflow that replaced manual deep package analysis and increased confidence when shipping code with new dependencies. | 高 | SU005, SU003 |
| CU011 | Replit says Socket reduces false positives and supports compliance work through integration with Vanta. | 中 | SU005, SU026 |
| CU012 | Vercel adopted Socket to manage dependency sprawl in a large monorepo, valued pnpm support, and worked with Socket on phased rollout features. | 高 | SU006, SU003 |
| CU013 | Cedar chose Socket after years of evaluating alternatives because earlier tools produced high alert volume, weak signal quality, and developer trust problems. | 中 | SU008 |
| CU014 | Cedar reports a 70% alert reduction, with workload falling from roughly 30 to 40 tickets per month to 10 to 12 Socket alerts per month. | 中 | SU008 |
| CU015 | Chia says about 90% of its security work now happens inside GitHub and that open security alerts across tools are down 70% after adopting Socket. | 高 | SU007, SU003 |
| CU016 | JumpCloud rolled Socket across more than 600 repositories and uses it for reachability, license management, SBOM support, and developer-endpoint protection. | 中 | SU010 |
| CU017 | Render says Socket has remained in its pull-request workflow for years because the alerts are actionable enough not to get removed as spam. | 中 | SU011 |
| CU018 | JupiterOne says Socket replaced multiple prior tools, reduced false positives through reachability, and fit a CI/CD-enforced security model after only a few hours of integration work. | 中 | SU015 |
| CU019 | Doctolib says Socket filled an automated supply-chain detection gap and was specifically valued when explaining security posture to external auditors. | 高 | SU009, SU003 |
| CU020 | Drata chose Socket to go beyond CVE-only tools and highlighted straightforward GitHub App deployment plus AI-detected supply-chain risk coverage. | 中 | SU012 |
| CU021 | MetaMask uses Socket alongside LavaMoat to identify suspicious packages early and relies on Socket's dependency search for very large JavaScript dependency trees. | 中 | SU013 |
| CU022 | SHI says Socket saved hundreds of engineer-hours and delivered an estimated 400 to 500 percent return on investment while fitting strict minimal-access requirements. | 中 | SU014 |
| CU023 | Public proof is much deeper for Anthropic, Replit, Vercel, Cedar, Chia, JumpCloud, Render, Doctolib, Drata, MetaMask, SHI, and JupiterOne than for xAI, Cursor, Figma, Gusto, and Mercado Libre, which are logo-only in the reviewed corpus. | 中 | SU001, SU002, SU003, SU004, SU005, SU006, SU007, SU008, SU009, SU010, SU011, SU012, SU013, SU014, SU015 |
| CU024 | Socket says it grew from 7,500 organizations after Series B to more than 27,000 by May 2026, indicating rapid breadth expansion during the AI-driven development cycle. | 高 | SU003, SU020 |
| CU025 | Socket says the Axios compromise drove more than 2,000 organizations to onboard within 24 hours, showing event-driven customer acquisition during acute supply-chain incidents. | 高 | SU002, SU003, SU019 |
| CU026 | Socket's public footprint disclosures describe organizations protected, repositories, and commits rather than paid-customer count or segment revenue mix. | 中 | SU003, SU020 |
| CU027 | Independent reviews generally praise Socket for behavioral analysis, GitHub PR integration, and free open-source access. | 中 | SU017, SU018, SU024, SU025 |
| CU028 | Independent reviews also warn that Socket is still maturing, is strongest in npm or JavaScript-heavy environments, and works best alongside a traditional CVE scanner rather than as a full replacement. | 中 | SU017, SU024, SU025 |
| CU029 | A January 2025 independent Medium test reported Java dependencies that failed to appear in Socket's UI or PR comments even after support acknowledged and partially fixed one issue. | 中 | SU016 |
| CU030 | Socket's own Vanta documentation says OAuth tokens are often revoked, which can make compliance synchronization appear broken until the integration is re-authorized. | 中 | SU026 |
| CU031 | Public case studies suggest Socket expands from PR-time scanning into API approvals, reachability, license and SBOM workflows, CI/CD gating, Vanta synchronization, dependency search, and developer-endpoint protection. | 高 | SU004, SU010, SU013, SU015, SU026, SU003 |
| CU032 | Many customer stories describe lean security teams embedding Socket into existing GitHub workflows rather than standing up a large dedicated AppSec operations function. | 中 | SU005, SU008, SU011, SU015 |
| CU033 | Customer testimonials emphasize lower noise and easier decision-making more often than direct hard-dollar savings, implying workflow quality is Socket's clearest public value proposition. | 中 | SU005, SU008, SU011, SU015 |
| CU034 | Reviewed public references span AI labs, developer tools, healthcare and regulated SaaS, identity, crypto/web3, and enterprise technology groups, but broad non-tech vertical proof remains limited. | 中 | SU001, SU003, SU008, SU009, SU013, SU014 |
| CU035 | Reviewed public materials do not disclose NRR, GRR, gross churn, contract length, or renewal cohorts for Socket customers. | 中 | SU001, SU003, SU017, SU024 |
| CU036 | Reviewed public materials do not disclose top-customer revenue concentration or the share of revenue tied to AI-native customers. | 中 | SU001, SU003, SU020 |
| CU037 | Because Socket's best-known references include Anthropic, Replit, Vercel, Cursor, xAI, and Figma, its customer brand appears unusually strong with AI-native engineering organizations. | 中 | SU002, SU003, SU019, SU020 |
| CU038 | GitHub-centric deployment and npm/JavaScript strength are clear product advantages, but the same pattern can limit confidence in broader heterogeneous enterprise environments until more ecosystem proof is public. | 中 | SU016, SU017, SU023, SU024, SU025 |
| CU039 | External news coverage largely repeats Socket's customer names and platform metrics rather than disclosing procurement detail, retention cohorts, or customer economics. | 中 | SU019, SU020, SU021, SU022 |
| CU040 | Even without formal retention metrics, Render's multi-year PR usage, JumpCloud's repo-wide integration, JupiterOne's CI/CD enforcement, and Replit/Doctolib compliance usage are favorable durability proxies. | 中 | SU005, SU009, SU010, SU011, SU015 |
| CU041 | GitHub Marketplace copy advertises five-minute deployment and inline PR feedback, corroborating the low-friction rollout described in customer case studies. | 中 | SU023 |
| CU042 | Open-source and community-heavy references such as Chia and MetaMask show Socket fits environments with large dependency trees, public contributors, or unusually high third-party code volume. | 中 | SU007, SU013 |
| CR001 | Socket announced a $60 million Series C on 2026-05-20 at a $1 billion valuation, and public sources say total funding reached $125 million. | 高 | SR001, SR026, SR027 |
| CR002 | Socket publicly names Anthropic, xAI, Replit, Cursor, Figma, Vercel, Gusto, Mercado Libre, and Cribl among its customers. | 中 | SR001 |
| CR003 | Socket said in the Coana acquisition announcement that it protects 8,500+ organizations and 750,000+ repositories, secures 2+ million commits each month, and identifies 500+ supply chain attacks every week. | 中 | SR002, SR028 |
| CR004 | Socket acquired Coana in April 2025 to add static control-flow and call-graph reachability analysis to its platform. | 高 | SR002, SR025, SR028 |
| CR005 | Both Socket and Coana said the entire Coana team joined Socket and that product integration was already underway after closing. | 高 | SR002, SR025 |
| CR006 | Coana and Socket said reachability analysis can eliminate up to 80% of false positives compared with traditional SCA tools. | 中 | SR002, SR025, SR028 |
| CR007 | Socket’s pricing page says Team includes precomputed reachability that cuts 60% of CVE false positives automatically, while Enterprise markets full-application reachability that can eliminate up to 90% of irrelevant CVEs. | 中 | SR004 |
| CR008 | SecurityWeek reported that Socket uses AI-assisted analysis plus human verification to detect supply chain compromises and prioritize remediation. | 中 | SR027, SR001 |
| CR009 | Socket positions itself as broader than CVE scanning by claiming to detect malicious packages, typosquats, license issues, low-quality packages, and other supply-chain risks. | 中 | SR008, SR009, SR016, SR017 |
| CR010 | Socket’s security and pricing materials say it never uploads or modifies customer source code and instead relies on dependency snapshots such as manifests and lockfiles. | 高 | SR005, SR004, SR011 |
| CR011 | Socket’s pricing page says only dependency lists are sent to Socket’s service and that payment data is processed by Stripe rather than Socket’s own servers. | 中 | SR004 |
| CR012 | Socket’s known-issues page says Socket for GitHub skips private npm package dependencies unless the private package repository is separately enabled or restructured as a workspace. | 中 | SR011 |
| CR013 | Socket’s ecosystem support page shows uneven product maturity: GitHub Actions support has no reachability or autofix, while several other surfaces are beta, experimental, planned, or unsupported. | 中 | SR015, SR024 |
| CR014 | Socket’s ecosystem support page says Swift is CVE-only with full support still in progress, and several ecosystems such as Objective-C, Elixir/Erlang, Dart, and Julia remain unsupported. | 中 | SR015 |
| CR015 | Socket recommends uv for best Python accuracy because pip dependency resolution is non-deterministic and poetry lockfiles do not lock optional dependencies. | 中 | SR015 |
| CR016 | Socket’s GitHub Actions and GitLab pipeline guides require customer-managed API keys or tokens and CI secret configuration to run scans inside customer workflows. | 中 | SR018, SR019 |
| CR017 | Socket’s GitLab pipeline guide explicitly says protected variables are safer and suggests separate least-privilege tokens or CI_JOB_TOKEN for unprotected branches. | 中 | SR019 |
| CR018 | Socket’s public status API shows operational dependence on the Socket API, dashboard, website, package pages, and multiple language-analysis components. | 中 | SR024 |
| CR019 | The same status API shows Socket expanding into .NET, Ruby, Rust, GitHub Actions, Chrome, OpenVSX, and HuggingFace analysis, widening the service surface the company must maintain. | 中 | SR024, SR015 |
| CR020 | GitHub now bundles dependency graph, SBOM export, the GitHub Advisory Database, Dependabot alerts, malware alerts, dependency review, and artifact attestations inside its security stack. | 中 | SR029 |
| CR021 | GitHub says dependency review can run in pull requests and its action can fail checks or block merges when vulnerable packages are introduced. | 中 | SR030 |
| CR022 | GitHub says Dependabot alerts are broadly available but cannot catch every security issue and may lag the arrival of new advisories in the GitHub Advisory Database. | 中 | SR031 |
| CR023 | Snyk now offers reachability analysis using static analysis, AI techniques, and expert validation, so reachability is no longer unique to Socket after Coana. | 中 | SR032, SR002 |
| CR024 | Snyk’s reachability documentation says a NO PATH FOUND result does not prove a vulnerability is unreachable or unexploitable. | 中 | SR032 |
| CR025 | npm trusted publishing replaces long-lived npm publish tokens with OIDC-based short-lived credentials tied to specific CI/CD workflows. | 中 | SR034 |
| CR026 | npm provenance lets maintainers prove where a package was built and published, but npm explicitly says provenance does not guarantee the package contains no malicious code. | 中 | SR035, SR036 |
| CR027 | GitHub’s provenance writeup says supply-chain attackers increasingly compromise publishing credentials rather than source code, making provenance an auditability control rather than a complete prevention mechanism. | 中 | SR036, SR035 |
| CR028 | Because npm and GitHub are raising the baseline with trusted publishing, provenance, dependency review, and malware alerts, Socket’s moat increasingly depends on precision, policy, and workflow execution rather than pure feature novelty. | 中 | SR029, SR030, SR031, SR034, SR035, SR036 |
| CR029 | Socket’s privacy policy says the company collects logs, cookies, and support data, works with third-party providers, and may share data to meet law or governmental requests. | 中 | SR006 |
| CR030 | Socket’s privacy policy was last updated on 2022-02-07, creating a freshness gap relative to the much broader 2025-2026 product and integration surface now marketed publicly. | 中 | SR006, SR024, SR004 |
| CR031 | Socket’s public agreements page shows an Enterprise Software License Agreement 1.2.0 effective 2026-03-23 and a Free Terms of Service 2.1.0, but the fetched public text does not expose liability, indemnity, or warranty details. | 中 | SR007 |
| CR032 | Socket’s security page says reports are stored on AWS S3 and its web servers are hosted on Render, making both providers material to availability and data handling. | 中 | SR005 |
| CR033 | Socket’s Vanta integration stores a refresh token in organization settings and its docs warn that Vanta often revokes tokens for undocumented reasons, making the compliance workflow brittle. | 中 | SR023 |
| CR034 | Socket’s SSO and SCIM features are available only to Enterprise organizations or Enterprise-plan customers. | 中 | SR021, SR022 |
| CR035 | Socket’s pricing and integration docs show GitLab, Bitbucket, Azure DevOps, self-hosted repositories, SCIM, audit logs, IP restrictions, and uptime SLA are gated to Enterprise. | 中 | SR004, SR019, SR020, SR021, SR022, SR033 |
| CR036 | Cooley confirmed Socket’s Series C and prior Series B financings, but the cited public financing materials still do not disclose ARR, churn, NRR, or customer concentration. | 中 | SR026, SR001 |
| CR037 | SecurityWeek reported Socket had approximately 100 employees at the time of the Series C. | 中 | SR027 |
| CR038 | Socket’s getting-started guide says Socket for GitHub is the easiest and most powerful approach, signaling a strong GitHub-centered distribution and workflow orientation. | 中 | SR010, SR018 |
| CR039 | Socket’s customers page lists case studies across Cedar, JumpCloud, SHI, JupiterOne, Anthropic, Doctolib, Replit, Chia, MetaMask, Drata, and Vercel. | 中 | SR003 |
| CR040 | Those public customer materials emphasize alert reduction, visibility, and workflow efficiency anecdotes rather than cohort retention, contract duration, or concentration metrics. | 中 | SR003, SR001 |
| CR041 | The European Commission says EU personal data may flow freely only to US companies participating in the Data Privacy Framework, while other GDPR transfer tools still require safeguards. | 中 | SR037 |
| CR042 | The EU Cyber Resilience Act entered into force on 2024-12-10, with reporting obligations beginning on 2026-09-11 and the main obligations applying from 2027-12-11. | 中 | SR038 |
| CR043 | The FTC cases database is a current monitoring venue for US enforcement, and this review did not identify a Socket-specific FTC matter there as of 2026-05-24. | 低 | SR039 |
| CR044 | Socket’s security page and pricing materials present no-source-code analysis, SOC 2 Type II posture, and enterprise controls as public mitigants, but those mitigants do not remove the need for alert precision or fresher privacy/legal documentation. | 中 | SR004, SR005, SR021, SR022 |
| CR045 | Socket’s security page centers founder Feross Aboukhadijeh, named security advisors, and security-industry investors as credibility anchors, which helps trust but also highlights founder-centric concentration. | 中 | SR005 |
| CR046 | The clearest monitorable thesis-break triggers are GHAS and GitHub-native displacement, rising alert noise despite Coana reachability, Coana-team integration slippage, stale privacy/legal docs, and continued absence of durability metrics into the next financing cycle. | 中 | SR002, SR029, SR030, SR032, SR006, SR026 |
| CR047 | Socket’s docs enumerate alert classes spanning malware, typosquats, Git and HTTP dependencies, telemetry, protestware, license, maintenance, and quality issues, which broadens coverage but also increases tuning burden and the chance of customer disagreement over noise. | 中 | SR013, SR014, SR016 |
| CR048 | Socket’s alert-action and policy controls partially mitigate noise through block, warn, monitor, and ignore workflows, but those controls still require ongoing customer configuration and integration upkeep. | 中 | SR012, SR023 |
| CV001 | Socket and SecurityWeek both reported that Socket raised $60 million in Series C funding at a $1 billion valuation in May 2026. | 高 | SV001, SV002 |
| CV002 | The May 2026 round was led by Thrive Capital with participation from a16z, Abstract Ventures, and Capital One Ventures. | 高 | SV001, SV031 |
| CV003 | SecurityWeek and Socket’s homepage indicate that Socket had raised about $125 million in total by May 2026. | 高 | SV002, SV004 |
| CV004 | The SaaS News said the Series C proceeds are intended to expand enterprise adoption and strengthen protection against AI-driven security threats. | 中 | SV031 |
| CV005 | Socket’s press release and syndicated coverage list Anthropic, xAI, Replit, Cursor, Figma, Vercel, Gusto, Mercado Libre, and Cribl among customers, alongside Fortune 100 enterprises. | 中 | SV001, SV003 |
| CV006 | Socket’s homepage says it protects more than 27,000 organizations. | 中 | SV004 |
| CV007 | Socket’s homepage says it protects more than 300,000 code repositories. | 中 | SV004 |
| CV008 | Socket’s homepage says 1.5 million developers trust the platform. | 中 | SV004 |
| CV009 | Socket’s homepage says it secures 11.6 million commits every month. | 中 | SV004 |
| CV010 | Socket’s homepage says it blocks more than 10,000 attacks every week. | 中 | SV004 |
| CV011 | Socket’s Latio market-report recap said supply-chain malware and securing AI-generated code accounted for 84% of practitioners’ top 2026 concerns. | 中 | SV008 |
| CV012 | Socket’s May 2025 Coana announcement said Socket revenue had more than tripled over the prior year. | 中 | SV009 |
| CV013 | Socket’s May 2025 Coana announcement said the company then protected 8,500+ organizations, 750,000+ repositories, and identified 500+ supply-chain attacks every week. | 中 | SV009 |
| CV014 | Socket’s 2026 research category shows frequent multi-ecosystem publication of supply-chain attack investigations across npm, Go, NuGet, RubyGems, PHP, and OpenVSX. | 中 | SV006 |
| CV015 | Socket’s 2025-2026 product category shows expansion into Jira, AI-agent skills scanning, Composer/PHP, immutable scans, OpenVSX, Ruby reachability, and Docker Hardened Images. | 中 | SV007 |
| CV016 | Socket and Coana said Coana’s reachability technology can eliminate 80%+ of false positives and drive up to 10x faster remediation. | 中 | SV009, SV010 |
| CV017 | Socket’s pricing page lists Team at $25 per developer per month, Business at $50 per developer per month, and Enterprise as custom priced. | 中 | SV005 |
| CV018 | Socket’s pricing page says Enterprise full-application reachability can eliminate up to 90% of irrelevant CVEs. | 中 | SV005 |
| CV019 | Socket’s FAQ says the product is free for open-source repositories but paid for private repositories beyond the first. | 中 | SV012 |
| CV020 | Socket’s GitHub Marketplace listing says the product supports 70+ red flags and detections across six categories. | 中 | SV014 |
| CV021 | Socket’s docs say customers include Vercel, Replit, and Brave, and that Next.js, Storybook, and MetaMask use Socket in open source. | 中 | SV011 |
| CV022 | GitHub Advanced Security says GitHub bundles static analysis, software composition analysis, and secret scanning directly into native GitHub workflows and explicitly contrasts that with third-party AppSec products. | 中 | SV016 |
| CV023 | GitHub Copilot says it has millions of users, tens of thousands of business customers, and can make developers up to 55% more productive. | 中 | SV017 |
| CV024 | JFrog’s market capitalization was $8.96 billion in May 2026. | 中 | SV018 |
| CV025 | JFrog’s Q1 2026 results showed $154.0 million of revenue, 26% year-over-year growth, and full-year 2026 revenue guidance of $628 million to $632 million. | 中 | SV019 |
| CV026 | JFrog’s Q1 2026 results also showed 80 customers above $1 million ARR, 1,225 customers above $100,000 ARR, and 120% trailing net dollar retention. | 中 | SV019 |
| CV027 | JFrog’s May 2026 market-cap-to-revenue proxy was about 14.2x using the midpoint of FY2026 revenue guidance. | 中 | SV018, SV019 |
| CV028 | GitLab’s market capitalization was $4.51 billion in May 2026. | 中 | SV020 |
| CV029 | GitLab’s fiscal 2026 Form 10-K reported $955.2 million of revenue, 26% growth, 87% gross margin, and 24% operating cash-flow margin. | 中 | SV021 |
| CV030 | GitLab’s May 2026 market-cap-to-revenue proxy was about 4.7x. | 中 | SV020, SV021 |
| CV031 | CrowdStrike’s market capitalization was $168.87 billion in May 2026. | 中 | SV022 |
| CV032 | CrowdStrike’s fiscal 2026 filing reported total revenue of $4.812 billion. | 中 | SV023 |
| CV033 | CrowdStrike’s May 2026 market-cap-to-revenue proxy was about 35.1x. | 中 | SV022, SV023 |
| CV034 | SentinelOne’s market capitalization was $6.38 billion in May 2026. | 中 | SV024 |
| CV035 | SentinelOne’s fiscal 2026 Form 10-K reported $1.0013 billion of revenue and 22% year-over-year growth. | 中 | SV025 |
| CV036 | SentinelOne’s May 2026 market-cap-to-revenue proxy was about 6.4x. | 中 | SV024, SV025 |
| CV037 | Palo Alto Networks’ market capitalization was $211.33 billion in May 2026. | 中 | SV026 |
| CV038 | Palo Alto Networks’ fiscal 2025 Form 10-K reported $9.2 billion of revenue and 14.9% growth. | 中 | SV027 |
| CV039 | Palo Alto Networks’ May 2026 market-cap-to-revenue proxy was about 23.0x. | 中 | SV026, SV027 |
| CV040 | Wiz’s official 2024 funding announcement said it raised $1 billion at a $12 billion valuation. | 中 | SV028 |
| CV041 | CNBC estimated Wiz ARR at about $350 million in 2024 and said the company counted 40% of Fortune 100 companies as customers. | 中 | SV029 |
| CV042 | Wiz’s disclosed valuation implied roughly 34.3x ARR. | 中 | SV028, SV029 |
| CV043 | GeekWire reported that Chainguard raised $356 million at a $3.5 billion valuation in April 2025, reached $40 million ARR after 7x growth, targeted more than $100 million ARR before fiscal 2026, and served 150+ customers. | 中 | SV030 |
| CV044 | Chainguard’s disclosed valuation implied about 87.5x current ARR or roughly 35x its near-term ARR target, making it a hypergrowth outlier comp. | 中 | SV030 |
| CV045 | A $1 billion valuation would require about $166.7 million ARR at 6x, $100 million at 10x, $66.7 million at 15x, $50 million at 20x, and $28.6 million at 35x revenue multiples. | 中 | SV018, SV019, SV020, SV021, SV022, SV023, SV024, SV025, SV026, SV027, SV028, SV029 |
| CV046 | With 1.5 million protected developers and a $35 blended monthly seat price between Team and Business, every 1% of monetized developers implies about $6.3 million of ARR. | 中 | SV004, SV005 |
| CV047 | Supporting a $1 billion valuation at a 20x multiple would require roughly $50 million of ARR, equivalent to about 8% monetization of the disclosed developer base at a $35 blended monthly seat price. | 中 | SV004, SV005 |
| CV048 | Because Socket is free for open-source repositories and has a $0 entry tier, disclosed protected-developer and protected-organization counts are only loose revenue proxies and likely overstate paid-seat volume. | 中 | SV005, SV012 |
| CV049 | The public comp bridge supports a $1 billion valuation only if Socket already monetizes closer to premium-private or high-teens public-devtools multiples rather than mature public-security multiples. | 中 | SV018, SV019, SV020, SV021, SV024, SV025, SV026, SV027, SV028, SV029, SV030 |
| CV050 | Socket’s Coana acquisition, reachability claims, and rapid 2025-2026 product expansion support a strategic premium above a plain SCA point solution. | 中 | SV007, SV009, SV010 |
| CV051 | GitHub’s native AppSec bundling and Copilot-led workflow control create real platform risk that argues against paying top-of-range private multiples without retention and ARR proof. | 中 | SV016, SV017 |
| CV052 | Public evidence shows strong top-of-funnel and enterprise credibility, but it does not disclose paid-customer count, ARR, NRR, gross margin, or cash burn. | 中 | SV001, SV004, SV005, SV012 |
| CV053 | A bear valuation range of about $450 million to $700 million is consistent with ARR landing around $25 million to $35 million and the market applying roughly 15x to 20x revenue. | 中 | SV004, SV005, SV020, SV021, SV024, SV025 |
| CV054 | A base valuation range of about $800 million to $1.1 billion is consistent with ARR around $45 million to $60 million and multiples around 18x to 22x. | 中 | SV004, SV005, SV018, SV019, SV026, SV027 |
| CV055 | A bull valuation range of about $1.2 billion to $1.7 billion requires ARR around $65 million to $85 million plus premium treatment closer to Wiz and Chainguard than to GitLab or SentinelOne. | 低 | SV028, SV029, SV030, SV020, SV021, SV024, SV025 |
| CV056 | A probability-weighted view centered in the high-$800 million to low-$900 million range makes the May 2026 round defensible but still somewhat ahead of what public evidence alone proves. | 中 | SV004, SV005, SV018, SV019, SV020, SV021, SV024, SV025, SV028, SV029, SV030 |
| CV057 | The right investment recommendation is track / research-more with conditional interest rather than an unconditional buy. | 中 | SV001, SV004, SV005, SV016, SV017 |
| CV058 | The best valuation stance is fair only if current ARR is already in the mid-tens of millions with strong retention; on disclosed public evidence alone the round reads slightly stretched. | 中 | SV004, SV005, SV018, SV019, SV020, SV021, SV024, SV025, SV028, SV029 |
| CV059 | Thesis-break triggers include ARR materially below about $40 million, low-single-digit paid conversion of the disclosed developer base, or evidence that GitHub bundling is slowing enterprise expansion. | 中 | SV004, SV005, SV016, SV017 |
| CV060 | No public source reviewed discloses Socket’s fully diluted cap table, liquidation preferences, or any secondary-liquidity terms. | 中 | SV001, SV002, SV004, SV005 |
| CV061 | The final diligence package should prioritize ARR, NRR, paid-versus-free conversion, enterprise ACV, burn efficiency, and cap-table terms. | 中 | SV005, SV012, SV016 |
| CV062 | The most supportable exit logic from public evidence is a later strategic sale or continued private scaling rather than a near-term IPO. | 低 | SV009, SV016, SV017 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Socket | Socket - Block zero-day supply chain attacks | Socket blocks malicious packages before they reach your code. |
| SO002 | Socket | Redefining Supply Chain Security - Socket | Founded in 2021, Socket offers a developer-first platform that proactively detects and blocks malicious packages in real time. |
| SO003 | Socket | Careers - Socket | We're on a mission to secure the world's software supply chains. |
| SO004 | Socket | Socket secures $40M to combat next-generation software supply chain attacks | San Francisco, CA — October 22, 2024. |
| SO005 | Socket | Socket raises $60M Series C at $1B valuation led by Thrive Capital | Today we're announcing Socket's $60 million Series C at a $1 billion valuation, led by Thrive Capital. |
| SO006 | Socket | Socket Raises $60M Series C at a $1B Valuation to Help Enterprises Build Securely With AI | Socket today announced it has raised $60 million in Series C funding at a $1 billion valuation. |
| SO007 | Socket | Socket Acquires Coana to Bring Reachability Analysis to Every AppSec Team | Today, we’re announcing a big step in securing the open source supply chain: Socket is acquiring Coana. |
| SO008 | Socket | Socket Has Acquired Secure Annex - Socket | Socket has acquired Secure Annex to expand extension security across browsers, IDEs, and AI tools. |
| SO009 | Socket | Introducing Socket Firewall: Free, Proactive Protection for Your Software Supply Chain | Socket Firewall Free: a lightweight tool that protects developer machines in real time, blocking malicious dependencies before they ever reach your laptop or build system. |
| SO010 | Socket | Pricing - Socket | Your source code never leaves your computer or your CI environment. Only your list of dependencies are sent to Socket's service. |
| SO011 | Socket | Getting started with Socket | Socket customers include Vercel, Replit, and Brave. |
| SO012 | Socket | Socket Firewall Overview | Socket Firewall is a suite of security tools that protects your development environment from malicious packages in real time. |
| SO013 | The SaaS News | Socket Raises $60M Series C at $1B Valuation | The SaaS News | Socket, a San Francisco, CA-based company offering a developer-first security platform, has raised $60 million in Series C funding at a $1 billion valuation. |
| SO014 | N2K CyberWire | Socket raises $60 million in Series C funding. | San Francisco-based software supply chain security company Socket has raised $60 million in Series C funding led by Thrive Capital. |
| SO015 | AiThority | Socket Raises $60M Series C at a $1B Valuation to Help Enterprises Build Securely With AI | Founded in 2020, Socket counts Anthropic, xAI, Replit, Cursor, Figma, Vercel, Gusto, Mercado LIbre, and Cribl among its customers. |
| SO016 | Signalbase | Socket Secures $60.0M | Socket just raised $60M Series C at a $1B valuation. Thrive Capital led ... $125M total funding. |
| SO017 | Tech Funding News | The startup that catches malicious code in minutes just raised $60M, hit a $1B valuation and enterprises are paying attention | Socket has raised $60M in a Series C round led by Thrive Capital, pushing the San Francisco-based company to a $1 billion valuation. |
| SO018 | Coana | Coana Joins Socket to Lead the Next Generation of AppSec | We are excited to announce that Coana has been acquired by Socket! |
| SO019 | StartupHub.ai | Socket Acquires Coana to Strengthen Software Composition Analysis (SCA) Offering | Socket’s acquisition of Coana brings best-in-class reachability analysis to application security teams globally. |
| SO020 | Security Systems News | Socket acquires Coana | SAN FRANCISCO – Socket ... today announced it has acquired Coana. |
| SO021 | GitHub | feross - Overview | Founder + CEO of Socket (@SocketDev). Started @webtorrent and @standard. Stanford lecturer for Web Security. |
| SO022 | Feross.org | Home of Feross Aboukhadijeh | Feross Aboukhadijeh is a computer security researcher, teacher, web developer, designer ... |
| SO023 | GitHub | GitHub - SocketDev/socket-cli: Command-line interface for socket.dev security analysis | Socket CLI is the command-line interface to Socket.dev, letting you scan dependencies, audit packages, and gate installs from your terminal or CI. |
| SO024 | GitHub | GitHub - SocketDev/sfw-free: Wraps your package manager, preventing installation of malicious packages. | Socket Firewall Free is a lightweight tool that protects developer machines in real time, blocking malicious dependencies before they ever reach your laptop or build system. |
| SO025 | GitHub | False positive: "URL strings" alert on textlint domain-checking rule · Issue #1126 · SocketDev/socket-cli | Please consider either: Marking this as a false positive for this package. |
| SO026 | The Register | Socket will block it with free malicious package firewall | AI detection alone can result in false positives. |
| SM001 | Socket | Socket - Block zero-day supply chain attacks | |
| SM002 | Socket | Features - Socket | |
| SM003 | Socket | Pricing - Socket | |
| SM004 | Socket | Supply Chain Risk | |
| SM006 | National Institute of Standards and Technology | NIST Special Publication (SP) 800-218, Secure Software Development Framework (SSDF) Version 1.1: Recommendations for Mitigating the Risk of Software Vulnerabilities | |
| SM007 | Cybersecurity and Infrastructure Security Agency | Software Bill of Materials (SBOM) | CISA | |
| SM008 | National Institute of Standards and Technology | Executive Order 14028, Improving the Nation's Cybersecurity | |
| SM009 | Openwall | security - backdoor in upstream xz/liblzma leading to ssh server compromise | |
| SM010 | Apache Logging Services | Security :: Apache Logging Services | |
| SM011 | Black Duck | 2026 OSSRA Report: Open Source Security & Risk Analysis | |
| SM012 | Sonatype | Software Supply Chain Risks | 2026 Software Supply Chain Report | |
| SM014 | Veracode | Spring 2026 Threat Research: Key Trends in Software Supply Chain Security | Veracode | |
| SM015 | Mordor Intelligence | Software Composition Analysis Market Size, Share Research Report, 2031 | |
| SM017 | Verified Market Reports | Global Software Supply Chain Security Market Size, Growth Trends & Forecast 2026-2034 | |
| SM018 | Research and Markets | Application Security Market Report 2026 - Research and Markets | |
| SM019 | Mordor Intelligence | Application Security Market Size, Scope, Demand Report 2031 | |
| SM020 | Fortune Business Insights | Application Security Market Size, Share | Industry Forecast 2034 | |
| SM021 | GitHub | About Dependabot version updates - GitHub Docs | |
| SM022 | GitHub | GitHub Advanced Security · Built-in protection for every repository | |
| SM023 | GitLab | Dependency scanning | GitLab Docs | |
| SM024 | GitLab | Pricing | |
| SM025 | Snyk | Snyk Plans and pricing | Try for Free or from $25/month | Get a Custom Quote | Snyk | |
| SM026 | npm | npm-audit | npm Docs | |
| SM027 | OSV | OSV - Open Source Vulnerabilities | |
| SM029 | OWASP | OWASP Dependency-Check | OWASP Foundation | |
| SM030 | Dependency-Track | Dependency-Track | Software Bill of Materials (SBOM) Analysis | |
| SM031 | Renovate | Renovate Docs | |
| SM032 | OpenSSF | Open Source Security Foundation – Linux Foundation Projects | |
| SM033 | European Commission | Cyber Resilience Act | |
| SP001 | Socket | Pricing - Socket | |
| SP002 | Socket | Socket Firewall - Socket | |
| SP003 | Socket | Socket Reachability - Socket | |
| SP004 | Socket | Full Application Reachability | |
| SP005 | Snyk | Snyk Plans and pricing | Try for Free or from $25/month | Get a Custom Quote | Snyk | |
| SP006 | Snyk | Open Source Security Management | Open Source SCA Tool | Snyk | |
| SP007 | Mend.io | Check Our Pricing - Mend.io | |
| SP008 | Endor Labs | AURI | AI-Native Application Security Platform | Endor Labs | |
| SP009 | Endor Labs | Endor Labs vs Socket Comparison | Application Security | Endor Labs | Socket's package signals can feel opaque, and policies are difficult to adapt to different environments. |
| SP010 | JFrog | Xray Main - 2023 | |
| SP011 | JFrog | Pricing 2026 | |
| SP012 | FOSSA | Pricing & Plans - FOSSA | |
| SP013 | GitHub | GitHub Advanced Security · Built-in protection for every repository · GitHub | |
| SP014 | GitHub Docs | About Dependabot alerts - GitHub Docs | |
| SP015 | Apiiro | Platform | |
| SP016 | Chainguard | Chainguard Pricing | |
| SP017 | Aikido Security | Pricing | Aikido Security | |
| SP018 | Aikido Security | Aikido, The Unified Security Platform | Aikido Security | |
| SP019 | OX Security | Application Security Platform: Code to Cloud | OX Security | |
| SP020 | OX Security | Welcome to OX Security Platform | OX docs | |
| SP021 | Upwind | Upwind Security: Cloud & AI Security for the Realtime Era | |
| SP022 | Pixee | Best SCA Tools for 2026: 9 Tools Compared | |
| SP023 | AppSec Santa | 8 Best Socket Alternatives (2026) | AppSec Santa | Socket is the go-to tool for catching supply-chain attacks before they merge — but it is not a general-purpose SCA platform. |
| SP024 | SourceForge | Endor Labs vs. Socket Comparison | |
| SP025 | PeerSpot | Compare GitHub Advanced Security vs Snyk | |
| SI001 | Socket | Socket Raises $60M Series C at a $1B Valuation to Help Enterprises Build Securely with AI | Socket today announced it has raised $60 million in Series C funding at a $1 billion valuation. |
| SI002 | Socket | Socket raises $60M Series C at $1B valuation led by Thrive Capital | The round brings our total funding to $125 million. |
| SI003 | Socket | Pricing - Socket | Team $25 ... Business $50 ... Enterprise Custom. |
| SI004 | Socket | Redefining Supply Chain Security - Socket | Founded in 2021, Socket offers a developer-first platform that proactively detects and blocks malicious packages in real time. |
| SI005 | Socket | Careers - Socket | We use best-in-class salary benchmarking to ensure market competitive compensation. |
| SI006 | Socket | Socket - Block zero-day supply chain attacks | Orgs Protected 27,000+ ... Code Repositories Protected 1.5M ... Commits Secured Every Month 11.6M+. |
| SI007 | Socket | Socket Acquires Coana to Bring Reachability Analysis to Every AppSec Team | Socket is acquiring Coana to bring best-in-class reachability analysis to every appsec team. |
| SI008 | The SaaS News | Socket Raises $60M Series C at $1B Valuation | |
| SI009 | RegTech Analyst | Socket hits $1bn valuation with $60m Series C raise | |
| SI010 | Cooley | Socket Raises $60 Million Series C at $1 Billion Valuation | Cooley advised Socket ... on its $60 million Series C at a $1 billion valuation, bringing its total funding to $125 million. |
| SI011 | TechCrunch | Socket lands a fresh $40M to scan software for security flaws | |
| SI012 | Cooley | Socket Secures $40 Million Series B | Cooley advised Socket ... on its $40 million Series B financing, bringing its total funding to $65 million. |
| SI013 | IT News Online / GlobeNewswire | Socket secures $40M to combat next-generation software supply chain security attacks led by industry titans Abstract Ventures, Elad Gil, and a16z | This latest round brings Socket's total funding to $65M. |
| SI014 | StartupHub.ai | Socket Secures $40M Series B to Safeguard Software Supply Chains Attacks | |
| SI015 | Tracxn | Socket company profile | Socket has raised $125M in funding. |
| SI016 | ZoomInfo | Socket - Overview, News & Similar companies | Revenue $18.1 Million. |
| SI017 | Forbes | Socket Acquires Coana To Build Out Its SCA Capabilities | Today’s announcement marks the next phase in the company’s development ... the deal with Coana – for an undisclosed sum. |
| SI018 | Business Partner Magazine | Socket Acquires Coana In Game-Changing Move For Cybersecurity Industry | The news comes as Socket has seen over 300% year-over-year revenue growth over the past year. |
| SI019 | Tech Funding News | Socket acquires Sequoia-backed Coana: 3 things to know about this game-changer in cybersecurity | While the acquisition price remains undisclosed to TFN, market analysts estimate it between $50 million and $100 million. |
| SI020 | Security Systems News | Socket acquires Coana | |
| SI021 | INCUBA | Coana writes a new chapter: Aarhus cyber startup becomes part of US Socket | The deal marks one of the largest exits in the INCUBA environment to date. |
| SI022 | OpenCorporates | Socket, Inc. company profile (Delaware) | |
| SI023 | Crunchbase | Socket - Crunchbase Company Profile & Funding | |
| SI024 | PitchBook | Socket company profile | |
| SI025 | Scamadviser | socket.dev Reviews | check if the site is a scam or legit | Trust Score 0 ... In summary, socket.dev is very likely not a scam but legit and reliable. |
| SI026 | Ashby | Socket Jobs | |
| SE001 | Socket | Socket - Block zero-day supply chain attacks | Socket blocks malicious packages before they reach your code. |
| SE002 | Socket | Getting started with Socket | |
| SE003 | Socket | Socket FAQ | In total, we look for 70+ signals in open source packages, which use different combinations of these 3 techniques – static analysis, package metadata analysis, and maintainer behavior analysis. |
| SE004 | Socket | Socket for GitHub - Socket | |
| SE005 | GitHub Marketplace | Socket Security - GitHub Marketplace | Socket currently supports 70 detections in 6 categories: Supply chain risk, Vulnerability, Quality, Maintenance, License, and more. |
| SE006 | Socket | Guide to Socket for VS Code | The Socket VS Code Extension is available in the VS Code extension marketplace and OpenVSX registry. |
| SE007 | Socket | Socket Firewall - Socket | Works across JavaScript, Python, Rust, and more with Enterprise support for Maven, Ruby, NuGet, and beyond. |
| SE008 | Socket | Socket Reachability - Socket | Cut CVE noise by up to 90% with Socket's Reachability Analysis. |
| SE009 | Socket | Full Application Reachability | |
| SE010 | Socket | Pricing - Socket | No. Your source code never leaves your computer or your CI environment. Only your list of dependencies are sent to Socket's service. |
| SE011 | Socket | Socket Acquires Coana to Bring Reachability Analysis to Every AppSec Team | No source code access needed for this demo. It’s fast, private, and uses “precomputed reachability analysis” to remove alerts from unused transitive dependencies. |
| SE012 | Coana | Coana Joins Socket to Lead the Next Generation of AppSec | |
| SE013 | Socket | Blog: Product News and Updates - Socket | |
| SE014 | Socket | Blog: Research News and Updates - Socket | North Korea’s Contagious Interview Campaign Spreads Across 5 Ecosystems, Delivering Staged RAT Payloads. |
| SE015 | GitHub | GitHub API: SocketDev organization | |
| SE016 | GitHub | GitHub API: SocketDev/socket-cli | |
| SE017 | GitHub | GitHub API: SocketDev/socket-vscode | |
| SE018 | GitHub | GitHub API: SocketDev/socket-sdk-js | |
| SE019 | GitHub | GitHub API: SocketDev/socket-sdk-python | |
| SE020 | Socket | Socket raises $60M Series C at $1B valuation led by Thrive Capital | Since we closed our Series B in October 2024, Socket has grown from 7,500 organizations to more than 27,000. We protect 1.5 million repositories and secure over 11.6 million commits every month. |
| SE021 | Socket | Socket Raises $60M Series C at a $1B Valuation to Help Enterprises Build Securely With AI | Socket analyzes the behavior of open source dependencies before they enter an organization’s codebase. |
| SE022 | AppSec Santa | Socket Review 2026: Supply Chain Attack Detection | Socket takes a different approach to SCA by focusing on supply chain attacks. Instead of checking dependencies against CVE databases, it analyzes what packages actually do at the code level. |
| SE023 | Security Systems News | Socket acquires Coana | |
| SE024 | Tech Funding News | Socket acquires Sequoia-backed Coana: 3 things to know about this game-changer in cybersecurity — TFN | |
| SE025 | SecurityWeek | Socket Raises $60 Million at $1 Billion Valuation | |
| SE026 | The SaaS News | Socket Raises $60M Series C at $1B Valuation | The SaaS News | |
| SE027 | RegTech Analyst | Socket hits $1bn valuation with $60m Series C raise | |
| SE028 | Ry Walker Research | Socket.dev | Ry Walker Research | Weaknesses: Paid product (free tier limited). Primarily JavaScript/Python/Go ecosystems. False positives possible with behavioral analysis. |
| SE029 | Startupik | Socket.dev: Detecting Malicious Code in Dependencies - Startupik | Startup magazine | |
| SU001 | Socket | Customers - Socket | Read the case studies below to see how we've helped top companies protect their teams from supply chain attacks. |
| SU002 | Socket | Socket Raises $60M Series C at a $1B Valuation to Help Enterprises Secure AI-Generated Code | Socket counts Anthropic, xAI, Replit, Cursor, Figma, Vercel, Gusto, Mercado Libre, and Cribl among its customers. |
| SU003 | Socket | Socket raises $60M Series C at $1B valuation led by Thrive Capital | Since we closed our Series B in October 2024, Socket has grown from 7,500 organizations to more than 27,000. |
| SU004 | Socket | How Anthropic Is Scaling Supply Chain Security with Socket | The manual review process ... has been almost entirely eliminated, with a 95% reduction in the need for hands-on scrutiny of dependencies. |
| SU005 | Socket | Building Secure Code with Confidence: How Replit Uses Socket to Reduce False Positives and Manage Supply Chain Risks | We're not getting as many false positives as some other systems would provide, so we don't tend to find ourselves getting blocked. |
| SU006 | Socket | Vercel Optimizes Open Source Dependency Management with Socket: Reduced Sprawl, Improved Hygiene, and Faster Decision-Making | Socket helped us get over the hurdle of continuous manual analysis. |
| SU007 | Socket | Enhancing Security and Streamlining Processes: How Chia Achieved a 70% Reduction in Open Security Alerts with Socket | Our number of open security alerts in GitHub from across all tools is down 70 percent. |
| SU008 | Socket | Cedar Cuts Vulnerability Alerts by 70% with Socket: Building Developer Trust Through Better Data Quality | We get now on average maybe 10 to 12 Socket alerts per month ... as opposed to previously when we were auto-generating 30 to 40 tickets a month. |
| SU009 | Socket | Doctolib Partners with Socket to Automate Supply Chain Threat Detection | When explaining our security posture to external auditors, Socket was always appreciated. |
| SU010 | Socket | JumpCloud Gains Visibility into Open Source and Developer Threats with Socket | The core functionality that uses the GitHub app was super easy. |
| SU011 | Socket | How Render Enables Scalable AppSec with Socket | Socket's been in our PRs for years. That's a good sign. |
| SU012 | Socket | Raising the Bar: How Drata Fortified Supply Chain Security with Socket | Integrating Socket was remarkably straightforward, especially with its GitHub app. |
| SU013 | Socket | MetaMask Leverages Socket for Proactive Threat Detection and Simplified Dependency Management | Socket is doing a big chunk of work now to identify potential threats before they reach us. |
| SU014 | Socket | SHI Strengthens Supply Chain Security with Socket: Reducing Manual Work and Human Error | Socket has saved the team significant time. Huffman estimates a 400-500% return on investment based on time saved. |
| SU015 | Socket | JupiterOne Secures Immutable Infrastructure with Socket's Streamlined CI/CD Security | We pulled out all the old stuff, dropped in Socket, and verified a few edge cases. |
| SU016 | Medium | SCA is NOT a Commodity: Lessons from Testing Socket.dev | Dependencies may not even be parsed, leaving your SDLC exposed. |
| SU017 | AppSecSanta | Socket Review 2026: Supply Chain Attack Detection | Socket takes a different approach to SCA by focusing on supply chain attacks. |
| SU018 | Startupik | Socket.dev: Detecting Malicious Code in Dependencies | Major companies like Figma, Vercel, and Brave publicly use and recommend Socket.dev. |
| SU019 | Techstartups | AI security startup Socket hits $1B valuation after $60M raise to stop software supply chain attacks | Within 24 hours, more than 2,000 organizations had onboarded to its platform. |
| SU020 | Tech Funding News | The startup that catches malicious code in minutes just raised $60M, hit a $1B valuation and enterprises are paying attention | Socket has grown from 7,500 organizations to more than 27,000. |
| SU021 | RegTech Analyst | Socket hits $1bn valuation with $60m Series C raise | Socket is now the standard for supply chain security at the companies building the most consequential AI products in the world. |
| SU022 | Pulse 2.0 | Socket: $60 Million Series C Raised At $1 Billion Valuation To Help Enterprises Secure AI-Generated Code | The round will support Socket's next phase of growth as more organizations adopt AI across software development. |
| SU023 | GitHub | Socket Security on GitHub Marketplace | Five minute deployment – Just install a GitHub app and you're done. |
| SU024 | ToolRadar | Socket Reviews, Pricing & Alternatives (2026) | 4.6/5 across review platforms. |
| SU025 | MakerStack | Socket Review (2026) | Rating: 7.4/10 ... Best for: dev teams using npm/PyPI heavily. |
| SU026 | Socket | Socket Vanta integration | Vanta often revokes these tokens. |
| SR001 | Socket | Socket Raises $60M Series C at a $1B Valuation to Help Enterprises Secure AI-Driven Development | Socket today announced it has raised $60 million in Series C funding at a $1 billion valuation. |
| SR002 | Socket | Socket Acquires Coana to Bring Reachability Analysis to Every AppSec Team | The entire Coana team have now joined Socket. |
| SR003 | Socket | Customers | |
| SR004 | Socket | Pricing | Enterprise ... need full application function-level reachability — eliminating up to 90% of irrelevant CVEs. |
| SR005 | Socket | Security Policy | We never upload your source code. |
| SR006 | Socket | Privacy Policy | We work with third parties to provide some of our Services. |
| SR007 | Socket | Terms of Service | Enterprise Software License Agreement 1.2.0 Effective: 23 March 2026. |
| SR008 | Socket | Socket vs Snyk | |
| SR009 | Socket | Socket vs Dependabot | |
| SR010 | Socket Docs | Getting started with Socket | |
| SR011 | Socket Docs | Known issues | Socket skips dependencies which are private npm packages. |
| SR012 | Socket Docs | Alert Actions and Triage Functionality | |
| SR013 | Socket Docs | Alert Types | |
| SR014 | Socket Docs | Alert Categories | |
| SR015 | Socket Docs | Ecosystem Support | GitHub Actions ... Reachability analysis ❌ ... Autofix ❌. |
| SR016 | Socket Docs | Supply Chain Risk | |
| SR017 | Socket Docs | Vulnerability | |
| SR018 | Socket Docs | Socket for GitHub Actions | The Action Workflow currently uses the auto generated GitHub Actions token. |
| SR019 | Socket Docs | Socket for GitLab Pipeline | Protected = safer ... Use $CI_JOB_TOKEN or a restricted-scope token for unprotected branches. |
| SR020 | Socket Docs | Socket for Azure DevOps | |
| SR021 | Socket Docs | SCIM | Available only to Enterprise organizations. |
| SR022 | Socket Docs | SSO (Single Sign-On) | SSO is available exclusively for customers on the Enterprise plan. |
| SR023 | Socket Docs | Vanta integration | Vanta often revokes these tokens. |
| SR024 | Socket Status | Status summary API | All Systems Operational. |
| SR025 | Coana | Coana Joins Socket to Lead the Next Generation of AppSec | Our entire team has joined Socket. |
| SR026 | Cooley | Socket Raises $60 Million Series C at $1 Billion Valuation | bringing its total funding to $125 million. |
| SR027 | SecurityWeek | Socket Raises $60 Million at $1 Billion Valuation | The company currently has approximately 100 employees. |
| SR028 | Tech Funding News | Socket acquires Sequoia-backed Coana: 3 things to know about this game-changer in cybersecurity | This acquisition follows Socket’s impressive 300% year-over-year revenue growth. |
| SR029 | GitHub Docs | GitHub security features | GitHub Code Security includes features that help you find and fix vulnerabilities, like code scanning, premium Dependabot features, and dependency review. |
| SR030 | GitHub Docs | About dependency review | By default, the dependency review action check will fail if it discovers any vulnerable packages. |
| SR031 | GitHub Docs | About Dependabot alerts | Alerts can’t catch every security issue. |
| SR032 | Snyk Docs | Reachability analysis | A vulnerability with the status NO PATH FOUND ... does not mean that the vulnerability is completely unreachable or unexploitable. |
| SR033 | GitLab Docs | Dependency scanning | |
| SR034 | npm Docs | Trusted publishing for npm packages | Trusted publishing allows you to publish npm packages directly from your CI/CD workflows using OpenID Connect (OIDC) authentication, eliminating the need for long-lived npm tokens. |
| SR035 | npm Docs | Generating provenance statements | When a package in the npm registry has established provenance, it does not guarantee the package has no malicious code. |
| SR036 | GitHub Blog | Introducing npm package provenance | Attackers instead attempt to inject malicious code into projects by directly compromising popular dependencies. |
| SR037 | European Commission | EU-US data transfers | Personal data can flow freely from the EU to companies in the United States that participate in the Data Privacy Framework. |
| SR038 | European Commission | Cyber Resilience Act | Reporting obligations [start] as of 11 September 2026. |
| SR039 | Federal Trade Commission | Cases and Proceedings | |
| SV001 | Socket | Socket Raises $60M Series C at a $1B Valuation to Help Enterprises Build Securely With AI | Socket today announced it has raised $60 million in Series C funding at a $1 billion valuation. |
| SV002 | SecurityWeek | Socket Raises $60 Million at $1 Billion Valuation | Supply chain protection provider Socket has announced raising $60 million in a Series C funding round that brings the total raised by the company to $125 million and its valuation to $1 billion. |
| SV003 | RegTech Analyst | Socket hits $1bn valuation with $60m Series C raise | Socket’s platform works by analysing the behaviour of open source dependencies before they are introduced into a codebase. |
| SV004 | Socket | Socket - Block zero-day supply chain attacks | Open source makes up 90% of modern application code. Socket scans every package and update for malicious behavior across all major registries. |
| SV005 | Socket | Pricing - Socket | Team $25 ... Business $50 ... Enterprise Custom. |
| SV006 | Socket | Research | Mini Shai-Hulud Hits @antv Ecosystem, 639 Compromised npm Package Versions. |
| SV007 | Socket | Product | Socket is now scanning AI agent skills across multiple languages and ecosystems, detecting malicious behavior before developers install, starting with skills.sh's 60,000+ skills. |
| SV008 | Socket | Socket Named a Supply Chain Innovator in Latio's 2026 AppSec Market Report | When practitioners were asked about their top concern for 2026, supply chain malware ranked among the top responses, alongside securing AI-generated code. Combined, those two categories accounted for 84% of responses. |
| SV009 | Socket | Socket Acquires Coana to Bring Reachability Analysis to Every AppSec Team | Socket revenue has more than tripled over the past year. |
| SV010 | Coana | Coana Joins Socket to Lead the Next Generation of AppSec | By applying reachability analysis to SCA, we enabled security teams to eliminate up to 80% of false positives compared to their traditional SCA tools. |
| SV011 | Socket Docs | Getting started with Socket | Socket customers include Vercel, Replit, and Brave. Socket is also used by prominent open source projects such as Next.js, Storybook, and Metamask. |
| SV012 | Socket Docs | Socket FAQ | Socket is free for open source repositories, forever. For private repositories beyond the first, Socket is paid. |
| SV013 | Socket | Socket for GitHub - Socket | Whenever a new dependency is added in a pull request, Socket analyzes the package's behavior and security risk. |
| SV014 | GitHub Marketplace | Socket Security - GitHub Marketplace | Socket currently supports 70 detections in 6 categories: Supply chain risk, Vulnerability, Quality, Maintenance, License, and more ... |
| SV015 | Socket Docs | Guide to Socket for VS Code | The extension only works on local files and does not integrate any organization-level settings like the GitHub App does. |
| SV016 | GitHub | GitHub Advanced Security | GitHub Advanced Security adds cutting-edge tools for static analysis, software composition analysis, and secret scanning to the GitHub platform that developers already know and love. |
| SV017 | GitHub | GitHub Copilot · Your AI pair programmer | Growing to millions of individual users and tens of thousands of business customers, GitHub Copilot is the world's most widely adopted AI developer tool. |
| SV018 | CompaniesMarketCap | JFrog (FROG) - Market capitalization | As of May 2026 JFrog has a market cap of $8.96 Billion USD. |
| SV019 | JFrog | JFrog Announces First Quarter 2026 Results | Revenue for the first quarter of 2026 was $154.0 million, up 26% year-over-year. |
| SV020 | CompaniesMarketCap | GitLab (GTLB) - Market capitalization | As of May 2026 GitLab has a market cap of $4.51 Billion USD. |
| SV021 | U.S. Securities and Exchange Commission | GitLab Form 10-K for fiscal year ended January 31, 2026 | We generated revenue of $955.2 million and $759.2 million in fiscal year 2026 and fiscal year 2025, respectively, representing growth of 26%. |
| SV022 | CompaniesMarketCap | CrowdStrike (CRWD) - Market capitalization | As of May 2026 CrowdStrike has a market cap of $168.87 Billion USD. |
| SV023 | U.S. Securities and Exchange Commission | CrowdStrike Form 10-K for fiscal year ended January 31, 2026 | Total revenue 4,812,005. |
| SV024 | CompaniesMarketCap | SentinelOne (S) - Market capitalization | As of May 2026 SentinelOne has a market cap of $6.38 Billion USD. |
| SV025 | U.S. Securities and Exchange Commission | SentinelOne Form 10-K for fiscal year ended January 31, 2026 | Our revenue was $1,001.3 million, $821.5 million, and $621.2 million for fiscal 2026, 2025, and 2024, respectively, representing year-over-year growth of 22% and 32%, respectively. |
| SV026 | CompaniesMarketCap | Palo Alto Networks (PANW) - Market capitalization | As of May 2026 Palo Alto Networks has a market cap of $211.33 Billion USD. |
| SV027 | U.S. Securities and Exchange Commission | Palo Alto Networks Form 10-K for fiscal year ended July 31, 2025 | For fiscal 2025 and 2024, total revenue was $9.2 billion and $8.0 billion, respectively, representing year-over-year growth of 14.9%. |
| SV028 | Wiz | Celebrating Our $1 Billion Funding Round and $12 Billion Valuation | Wiz has raised $1 billion at a $12 billion valuation. |
| SV029 | CNBC | Wiz: 2024 CNBC Disruptor 50 | The New York-based company with Israeli roots has roughly tripled its annual recurring revenue over the past two years to an estimated $350 million. |
| SV030 | GeekWire | Cybersecurity startup Chainguard lands $356M at $3.5B valuation, up from $1.1B a year ago | Chainguard said it grew annual recurring revenue 7X to $40 million in its fiscal year 2025, and plans to reach more than $100 million in ARR before fiscal year 2026. |
| SV031 | The SaaS News | Socket Raises $60M Series C at $1B Valuation | The company will use the funding to scale its software supply chain security platform, expand enterprise adoption, and strengthen protections against malicious open-source dependencies and AI-driven security threats. |