latio-secure-supply-chain-skills
Audit and take action to secure your repository against supply chain misconfigurations and malware.
Open source Repository Open in the app JSON README (API)
About
Audit and take action to secure your repository against supply chain misconfigurations and malware.
Details
- Kind
- Plugins
- Topic
- Security & identity
- Publisher
- latiotech
- Origin
- marketplace
- Category
- ferramentas
- Stars
- 15
- Forks
- 4
- Open pull requests
- 1
- Last push
- 2026-03-30T23:37:30Z
- Repository state
- ativo
- License
- MIT
- Added
- 2026-08-30 01:48:58
- Updated
- 2026-08-30 01:48:58
- Origin id
latiotech/secure-supply-chain-skills/latio-secure-supply-chain-skills
README
# Supply Chain Security Plugin for Claude Code Protect your projects from supply chain attacks with Claude. ## What Is This? When 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**. This plugin gives Claude Code the ability to **audit your project** for supply chain risks and **fix them for you**, covering seven areas: | Area | Examples of What Could Go Wrong | |------|-------------------------------| | **Packages** (npm, pip, etc.) | A dependency gets hijacked and runs a malicious install script | | **Containers** (Docker) | Your base image silently changes to include a backdoor | | **GitHub Actions** | A third-party Action gets compromised and steals your secrets | | **Infrastructure as Code** (Terraform) | An unverified module provisions resources you didn't ask for | | **AI/ML Models** | A pickle file executes arbitrary code when loaded | | **IDE Extensions** | A VS Code extension exfiltrates code from your workspace | | **Credentials** | A leaked API key or long-lived token enables lateral movement across services | Based on the [Latio](https://latio.com) Supply Chain Security Checklist ## Getting Started ### Prerequisites - [Claude Code](https://docs.anthropic.com/en/docs/claude-code) installed and working - [SSH access to GitHub](https://docs.github.com/en/authentication/connecting-to-github-with-ssh) configured (the plugin installer clones via SSH) ### Installation > **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`. Add the marketplace and install the plugin from within Claude Code: ``` /plugin marketplace add latiotech/secure-supply-chain-skills /plugin install supply-chain-security@secure-supply-chain ``` After installing, run `/reload-plugins` to activate the plugin. ### Your First Audit Once installed, open Claude Code in any project and run: ``` /audit-supply-chain ``` This 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. ## Two Types of Commands ### Action Commands - Fix Issues Directly These 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. | Command | What It Does | |---------|-------------| | `/audit-supply-chain` | Full audit - detects what's in your repo, checks everything, auto-fixes critical items | | `/harden-packages` | Pin dependency versions, disable install scripts, secure registry configs | | `/harden-containers` | Pin base images by digest, enforce non-root, create .dockerignore | | `/harden-actions` | Pin Actions to SHAs, set permissions, fix script injection, add Dependabot | | `/harden-iac` | Pin Terraform modules/providers, generate lockfiles, flag provisioners | | `/harden-ai-ml` | Fix unsafe torch.load/pickle.load, pin model sources, add hash verification | | `/harden-ide-extensions` | Audit extensions, remove secrets from settings, add devcontainer config | | `/harden-credentials` | Scan for leaked secrets, set up pre-commit hooks, harden .gitignore, fix credential anti-patterns | | `/audit-credentials` | Find long-lived tokens, hardcoded secrets, credentials that should be rotated or replaced with OIDC | | `/update-pins` | Check pinned deps, Actions, images, and modules for newer versions — auto-updates patch/minor, flags majors with changelogs | | `/minimize` | Remove unused dependencies and convert Dockerfiles to multi-stage builds to reduce attack surface | ### Walkthrough Commands - Guided Setup for Advanced Items These 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. | Command | What It Walks You Through | |---------|--------------------------| | `/setup-oidc` | Replace long-lived cloud credentials with OIDC tokens (AWS, GCP, Azure) | | `/setup-image-signing` | Sign container images with Cosign/Sigstore | | `/setup-tag-rulesets` | Protect release tags from force-push attacks (GitHub rulesets) | | `/setup-admission-control` | Enforce image policies in Kubernetes (Kyverno/OPA) | | `/setup-sbom` | Generate SBOMs and SLSA provenance attestations | | `/setup-commit-signing` | Set up commit and tag signing (SSH, GPG, or Sigstore gitsign) | | `/setup-runner-monitoring` | Add runtime detection to CI runners (StepSecurity/Falco) | ## How It Works 1. **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). 2. **Analysis** - For each relevant domain, checks are run ordered by severity. 3. **Action** - Critical issues are fixed directly: versions pinned, SHAs resolved, configs secured. Each change is explained. 4. **Report** - A summary shows what was fixed, what needs manual attention, and which walkthrough commands to run next. ## What Gets Checked and Fixed <details> <summary><strong>Packages</strong> - npm, pnpm, yarn, pip, uv, Go modules, Cargo, Bundler, Maven, Gradle</summary> - Are dependency versions pinned (no floating ranges)? - Are install scripts disabled where possible? - Is the lockfile present and up to date? - Are registry scopes configured to prevent dependency confusion? </details> <details> <summary><strong>Containers</strong> - Docker, OCI images</summary> - Are base images pinned by digest (not just a mutable tag)? - Do containers run as non-root? - Are multi-stage builds used to minimize the final image? - Are images signed and verified? </details> <details> <summary><strong>GitHub Actions</strong> - Workflows and third-party Actions</summary> - Are Actions pinned to full commit SHAs (not tags)? - Are workflow permissions set to least privilege? - Are `pull_request_target` triggers used safely? - Is CODEOWNERS configured for workflow files? </details> <details> <summary><strong>Infrastructure as Code</strong> - Terraform modules and providers</summary> - Are modules and providers version-pinned? - Is the lockfile committed? - Are dangerous provisioners (local-exec, remote-exec) flagged? - Is state encrypted and access-controlled? </details> <details> <summary><strong>AI/ML Models</strong> - Model files and loading code</summary> - Are pickle files present (arbitrary code execution risk)? - Does code use unsafe `torch.load()` or `pickle.load()`? - Are models verified by hash or signature? - Are safer serialization formats (safetensors, ONNX) used? </details> <details> <summary><strong>IDE Extensions</strong> - VS Code, devcontainers</summary> - Are recommended extensions from verified publishers? - Are secrets accidentally stored in workspace settings? - Is devcontainer extension isolation configured? - Is an extension allowlist in place? </details> <details> <summary><strong>Credentials</strong> - Secrets, tokens, API keys</summary> - Are there hardcoded secrets in source code or config files? - Are `.env` files properly gitignored? - Are secrets present in git history? - Is a pre-commit hook set up to prevent future leaks? - Are long-lived cloud credentials replaceable with OIDC? - Are package manager auth tokens secured? - Are commits and tags signed to prevent impersonation? </details> ## The Skill Beyond 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: - A prioritized checklist of "Do Right Now" vs "Do Eventually" items - A catalog of open-source and paid security tools for each domain - Ready-to-use configuration snippets for all supported package managers and tools Just ask Claude something like *"How do I prevent dependency confusion attacks?"* and it will pull from this knowledge automatically. ## Status and Contributing This 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: - **Open an issue**: [github.com/latiotech/secure-supply-chain-skills/issues](https://github.com/latiotech/secure-supply-chain-skills/issues) - **Submit a PR**: See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines Feedback from real-world usage is especially valuable — if a check produces a false positive, misses something, or breaks your workflow, let us know.