supply-chain-security – Security skill for AI agents
Software supply-chain security assessment: SBOM generation, SCA scanning, CI/CD pipeline review, container image audit, build integrity, dependency provenance, and vulnerability reachability verification.
Install the supply-chain-security skill with one command
Pick the tool you use, copy the command and paste it in your terminal. The skill lands in that tool's skills folder and is ready in your next session.
Claude Code
mkdir -p ~/.claude/skills && curl -fsSL https://skill.dn1nguoi.com/z/supply-chain-security-99084.tgz | tar xz -C ~/.claude/skills
Call it in Claude Code: /supply-chain-security – then describe the task you want done.
Codex
mkdir -p ~/.codex/skills && curl -fsSL https://skill.dn1nguoi.com/z/supply-chain-security-99084.tgz | tar xz -C ~/.codex/skills
Call it in Codex: $supply-chain-security – then describe the task you want done.
Hermes Agent
mkdir -p ~/.hermes/skills && curl -fsSL https://skill.dn1nguoi.com/z/supply-chain-security-99084.tgz | tar xz -C ~/.hermes/skills
Call it in Hermes Agent: /supply-chain-security – then describe the task you want done.
DeepSeek
mkdir -p ~/.deepcode/skills && curl -fsSL https://skill.dn1nguoi.com/z/supply-chain-security-99084.tgz | tar xz -C ~/.deepcode/skills
Call it in DeepSeek: /supply-chain-security – then describe the task you want done.
SKILL.md content
Supply Chain Security Testing
When to Use
- Auditing how software is built, packaged, and depended upon.
- Verifying whether a disclosed CVE is actually reachable in a project.
适用场景
- 软件供应链安全评估
- 开源依赖漏洞扫描与验证
- CI/CD 管道安全审计
- 容器镜像安全分析
- 第三方组件合规审查
- 构建产物溯源与完整性验证
六层供应链治理框架
Layer 1: 源码信任评估 → 上游仓库/维护者/发布历史审查
Layer 2: 构建管道集成 → CI/CD 安全门禁、签名验证
Layer 3: 制品分发完整性 → 签名、校验和、SBOM 附加
Layer 4: 运行时保护 → 容器扫描、准入控制
Layer 5: 持续监控 → CVE 实时追踪、漏洞可达性分析
Layer 6: 事件响应 → 供应链攻击应急、回滚策略工作流
1. SBOM 生成与审计
生成 SBOM:
□ CycloneDX 格式: cdxgen → bom.json
□ SPDX 格式: sbom-tool generate
□ Syft: syft <image|dir> -o spdx-json
审计要点:
□ 是否存在未知/未授权的依赖
□ 是否存在已废弃/停止维护的包
□ 许可证冲突检测
□ 直接依赖 vs 传递依赖清单
□ 每个组件的发布时间线和维护者状态2. 软件组成分析(SCA)
# OSV-Scanner(免费、Google 维护)
osv-scanner scan -r . --format json
# OWASP Dependency-Track(企业级持续监控)
# Deployment requires reviewed configuration, digest, access controls and authorization.
# → 上传 SBOM → 自动匹配 NVD/OSV/GitHub Advisory
# Snyk(商业)
snyk test --all-projects
# Persistent monitoring is a separately authorized operation.
# Trivy(容器 + 依赖 + IaC)
trivy fs . # 文件系统扫描
trivy image nginx # 容器镜像
trivy config . # IaC 配置3. 漏洞可达性验证
SCA 告警需要结合实际版本、暴露面和调用路径验证。未提供可复现样本时,不声称固定的可达比例。
验证步骤:
1. 用 Dependency-Track 或 Trivy 获取 CVE 列表
2. 筛选 CVSS ≥ 7.0 的漏洞
3. 对有 PoC 的 CVE 做可达性分析
- Code Property Graph 切片: 追踪用户输入到漏洞函数的路径
- DEPTEX 方法: EPD (Execution Path Dominance) + LLM 语义验证
4. 在隔离环境中验证 PoC
5. 对可达的漏洞按实际影响排序修复优先级工具参考:
- CodeQL: GitHub 代码查询 → 数据流分析
- Snyk Code: 可达性标记
- DEPTEX: LLM 辅助上下文感知风险评估
4. CI/CD 管道安全
安全检查点:
□ 代码提交 → pre-commit hook: gitleaks (密钥扫描)
□ PR 阶段 → SCA 扫描 (Trivy/OSV-Scanner)
□ 构建阶段 → 制品签名 (cosign)
□ 推送阶段 → SBOM 附加 (syft + attest)
□ 部署阶段 → 准入控制 (OPA/Kyverno + 镜像扫描)
□ 运行时 → 持续漏洞监控 (Dependency-Track)
管道自身安全:
□ Pipeline as Code 审计(GitHub Actions / GitLab CI 配置注入)
□ Runner 隔离(防止恶意构建突破容器)
□ 密钥管理(Actions Secrets / Vault,禁止硬编码)
□ 第三方 Action 审查(锁定 commit SHA,非 tag)5. 容器镜像安全
# Dockerfile 审计
hadolint Dockerfile
# 镜像扫描(多层:OS + 应用依赖 + 配置)
trivy image --severity HIGH,CRITICAL nginx:latest
# 最小基础镜像
# 优先: distroless → alpine → slim → 避免 latest
docker scout quickview nginx:latest
# 镜像签名
cosign sign --key cosign.key myimage:tag
cosign verify --key cosign.pub myimage:tag6. 第三方依赖审查
新增依赖 Checklist:
□ 维护状态:最近 6 个月有提交?维护者活跃度?
□ 安全历史:过去有无被植入恶意代码?
□ 依赖树:引入后新增多少传递依赖?
□ 许可证:与项目许可证兼容?
□ 替代方案:有无更安全的替代(Snyk Advisor / Socket.dev 评分)?
风险评估矩阵:
高维护 × 低依赖数 × 兼容许可证 → 低风险
低维护 × 高依赖数 × 许可证冲突 → 高风险工具链
使用项目已安装并锁定版本的 OSV-Scanner、Trivy、Syft、cdxgen、Cosign、Gitleaks 或 CodeQL。核对真实可执行路径及官方版本说明;安装新工具需先审核来源、不可变版本和校验和,不执行在线下载脚本或自动全局安装。工具能力、付费条件和可达性支持取决于实际版本,不能从工具名推断。
扫描私有依赖、上传 SBOM、持续监控、公开镜像拉取、签名及推送均有网络、数据或成本边界。先确认目标和授权;默认仅在获准的本地目录生成报告,不启动服务或注册监控。
参考
references/sbom-sca-methodology.md— SBOM + SCA 方法论references/cicd-pipeline-security.md— CI/CD 管道安全审计
任务完成自检(声称完成前 MUST 通过)
- [ ] 我是否执行了工作流中的每一步(而不是只阅读)?
- [ ] 我是否基于
tool-index使用了真实工具路径? - [ ] 我是否产出了可复现证据(命令/脚本/截图/报告)?
- [ ] 我是否完成并回写了 RULES 要求的 Checklist 项?
Limitations
- SBOM completeness depends on ecosystem tooling maturity.
- Reachability analysis is heuristic; manual confirmation advised.
> Adapted from [zhaoxuya520/reverse-skill](https://github.com/zhaoxuya520/reverse-skill) (MIT).
Related skills
- swiftui-performance-audit – Audit SwiftUI performance issues from code review and profiling evidence.
- threat-hunting – Blue-team threat hunting: detection engineering with Sigma/YARA, SIEM query design, and…
- threat-intelligence – Authorized OSINT and cyber threat intelligence: enriching IOCs, campaigns…
- threat-mitigation-mapping – Map identified threats to appropriate security controls and mitigations. Use when…
- threat-modeling – Conduct threat modeling using STRIDE methodology. Identify threats, assess risks, and…
- threat-modeling-expert – Expert in threat modeling methodologies, security architecture review, and risk…
- top-web-vulnerabilities – Provide an illustrative, overlapping reference for common web application vulnerability…
- usage-engineer – Cut wasted tokens in coding-agent sessions: compact long logs, stop identical retries…