{
  "markdown": "# Supply Chain Security Plugin for Claude Code\n\nProtect your projects from supply chain attacks with Claude.\n\n## What Is This?\n\nWhen you build software, you rely on hundreds of third-party packages, container images, GitHub Actions, and more. Attackers increasingly target these dependencies to sneak malicious code into otherwise trustworthy projects. This is called a **supply chain attack**.\n\nThis plugin gives Claude Code the ability to **audit your project** for supply chain risks and **fix them for you**, covering seven areas:\n\n| Area | Examples of What Could Go Wrong |\n|------|-------------------------------|\n| **Packages** (npm, pip, etc.) | A dependency gets hijacked and runs a malicious install script |\n| **Containers** (Docker) | Your base image silently changes to include a backdoor |\n| **GitHub Actions** | A third-party Action gets compromised and steals your secrets |\n| **Infrastructure as Code** (Terraform) | An unverified module provisions resources you didn't ask for |\n| **AI/ML Models** | A pickle file executes arbitrary code when loaded |\n| **IDE Extensions** | A VS Code extension exfiltrates code from your workspace |\n| **Credentials** | A leaked API key or long-lived token enables lateral movement across services |\n\nBased on the [Latio](https://latio.com) Supply Chain Security Checklist\n\n## Getting Started\n\n### Prerequisites\n\n- [Claude Code](https://docs.anthropic.com/en/docs/claude-code) installed and working\n- [SSH access to GitHub](https://docs.github.com/en/authentication/connecting-to-github-with-ssh) configured (the plugin installer clones via SSH)\n\n### Installation\n\n> **Note:** This plugin is pending addition to the official Claude Code marketplace. In the meantime, you can install it directly using the self-hosted marketplace below. Once it's in the official marketplace, you'll be able to install it with just `/plugin install supply-chain-security`.\n\nAdd the marketplace and install the plugin from within Claude Code:\n\n```\n/plugin marketplace add latiotech/secure-supply-chain-skills\n/plugin install supply-chain-security@secure-supply-chain\n```\n\nAfter installing, run `/reload-plugins` to activate the plugin.\n\n### Your First Audit\n\nOnce installed, open Claude Code in any project and run:\n\n```\n/audit-supply-chain\n```\n\nThis scans your repo, figures out which areas apply to you, **auto-fixes critical items** (pinning versions, resolving SHAs, disabling install scripts), and gives you a report showing what was fixed and what still needs attention.\n\n## Two Types of Commands\n\n### Action Commands - Fix Issues Directly\n\nThese commands **take action by default**. They scan your repo, make changes (pin versions to exact hashes, resolve commit SHAs, fix configs), and explain each change as it's made. Run these first.\n\n| Command | What It Does |\n|---------|-------------|\n| `/audit-supply-chain` | Full audit - detects what's in your repo, checks everything, auto-fixes critical items |\n| `/harden-packages` | Pin dependency versions, disable install scripts, secure registry configs |\n| `/harden-containers` | Pin base images by digest, enforce non-root, create .dockerignore |\n| `/harden-actions` | Pin Actions to SHAs, set permissions, fix script injection, add Dependabot |\n| `/harden-iac` | Pin Terraform modules/providers, generate lockfiles, flag provisioners |\n| `/harden-ai-ml` | Fix unsafe torch.load/pickle.load, pin model sources, add hash verification |\n| `/harden-ide-extensions` | Audit extensions, remove secrets from settings, add devcontainer config |\n| `/harden-credentials` | Scan for leaked secrets, set up pre-commit hooks, harden .gitignore, fix credential anti-patterns |\n| `/audit-credentials` | Find long-lived tokens, hardcoded secrets, credentials that should be rotated or replaced with OIDC |\n| `/update-pins` | Check pinned deps, Actions, images, and modules for newer versions — auto-updates patch/minor, flags majors with changelogs |\n| `/minimize` | Remove unused dependencies and convert Dockerfiles to multi-stage builds to reduce attack surface |\n\n### Walkthrough Commands - Guided Setup for Advanced Items\n\nThese commands are **interactive walkthroughs** for configurations that require steps outside your codebase (cloud provider setup, Kubernetes config, GitHub settings). Run these separately when you're ready to tackle advanced hardening.\n\n| Command | What It Walks You Through |\n|---------|--------------------------|\n| `/setup-oidc` | Replace long-lived cloud credentials with OIDC tokens (AWS, GCP, Azure) |\n| `/setup-image-signing` | Sign container images with Cosign/Sigstore |\n| `/setup-tag-rulesets` | Protect release tags from force-push attacks (GitHub rulesets) |\n| `/setup-admission-control` | Enforce image policies in Kubernetes (Kyverno/OPA) |\n| `/setup-sbom` | Generate SBOMs and SLSA provenance attestations |\n| `/setup-commit-signing` | Set up commit and tag signing (SSH, GPG, or Sigstore gitsign) |\n| `/setup-runner-monitoring` | Add runtime detection to CI runners (StepSecurity/Falco) |\n\n## How It Works\n\n1. **Detection** - Commands look at your repo's files to figure out which domains apply (e.g., if you have a `package.json`, it checks packages; if you have a `Dockerfile`, it checks containers).\n2. **Analysis** - For each relevant domain, checks are run ordered by severity.\n3. **Action** - Critical issues are fixed directly: versions pinned, SHAs resolved, configs secured. Each change is explained.\n4. **Report** - A summary shows what was fixed, what needs manual attention, and which walkthrough commands to run next.\n\n## What Gets Checked and Fixed\n\n<details>\n<summary><strong>Packages</strong> - npm, pnpm, yarn, pip, uv, Go modules, Cargo, Bundler, Maven, Gradle</summary>\n\n- Are dependency versions pinned (no floating ranges)?\n- Are install scripts disabled where possible?\n- Is the lockfile present and up to date?\n- Are registry scopes configured to prevent dependency confusion?\n</details>\n\n<details>\n<summary><strong>Containers</strong> - Docker, OCI images</summary>\n\n- Are base images pinned by digest (not just a mutable tag)?\n- Do containers run as non-root?\n- Are multi-stage builds used to minimize the final image?\n- Are images signed and verified?\n</details>\n\n<details>\n<summary><strong>GitHub Actions</strong> - Workflows and third-party Actions</summary>\n\n- Are Actions pinned to full commit SHAs (not tags)?\n- Are workflow permissions set to least privilege?\n- Are `pull_request_target` triggers used safely?\n- Is CODEOWNERS configured for workflow files?\n</details>\n\n<details>\n<summary><strong>Infrastructure as Code</strong> - Terraform modules and providers</summary>\n\n- Are modules and providers version-pinned?\n- Is the lockfile committed?\n- Are dangerous provisioners (local-exec, remote-exec) flagged?\n- Is state encrypted and access-controlled?\n</details>\n\n<details>\n<summary><strong>AI/ML Models</strong> - Model files and loading code</summary>\n\n- Are pickle files present (arbitrary code execution risk)?\n- Does code use unsafe `torch.load()` or `pickle.load()`?\n- Are models verified by hash or signature?\n- Are safer serialization formats (safetensors, ONNX) used?\n</details>\n\n<details>\n<summary><strong>IDE Extensions</strong> - VS Code, devcontainers</summary>\n\n- Are recommended extensions from verified publishers?\n- Are secrets accidentally stored in workspace settings?\n- Is devcontainer extension isolation configured?\n- Is an extension allowlist in place?\n</details>\n\n<details>\n<summary><strong>Credentials</strong> - Secrets, tokens, API keys</summary>\n\n- Are there hardcoded secrets in source code or config files?\n- Are `.env` files properly gitignored?\n- Are secrets present in git history?\n- Is a pre-commit hook set up to prevent future leaks?\n- Are long-lived cloud credentials replaceable with OIDC?\n- Are package manager auth tokens secured?\n- Are commits and tags signed to prevent impersonation?\n</details>\n\n## The Skill\n\nBeyond the slash commands, this plugin includes a `supply-chain-hardening` skill that activates automatically when you ask Claude about supply chain security topics. It gives Claude access to:\n\n- A prioritized checklist of \"Do Right Now\" vs \"Do Eventually\" items\n- A catalog of open-source and paid security tools for each domain\n- Ready-to-use configuration snippets for all supported package managers and tools\n\nJust ask Claude something like *\"How do I prevent dependency confusion attacks?\"* and it will pull from this knowledge automatically.\n\n## Status and Contributing\n\nThis plugin is a new initiative and actively evolving. Coverage, scanner integrations, and fix strategies will improve over time. If you run into issues, have ideas, or want to improve the checks:\n\n- **Open an issue**: [github.com/latiotech/secure-supply-chain-skills/issues](https://github.com/latiotech/secure-supply-chain-skills/issues)\n- **Submit a PR**: See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines\n\nFeedback from real-world usage is especially valuable — if a check produces a false positive, misses something, or breaks your workflow, let us know.",
  "bytes": 9049,
  "sha": "41836ae47c2340ebc4f7cee7e79c669a5e923150c261abf38a2fe22236ad3f9c",
  "repo_slug": "latiotech/secure-supply-chain-skills",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_latiotech_secure_supply_chain_skills_lat_a6c2c277/readme"
}