Supply Chain Risk Auditor Skill for Claude Code: What It Does and How to Install It

Makes Claude audit a project's npm, PyPI and Go dependencies for known advisories, abandoned upstreams, publisher concentration and install scripts.

Installs: 8.5K; on skills.sh
Stars: 7.9K; on GitHub
Files: 12; besides SKILL.md
Updated: Oct 2026; last repo commit

Claude runs two Python scripts, collect.py and render.py, and writes a Markdown supply-chain report. The scripts measure the facts, so Claude does not estimate maintainer counts, downloads or CVE history. Claude then adds a short narrative, upgrade paths and replacement ideas, kept separate from what the scripts measured.

It needs uv and a project with package.json, pyproject.toml, requirements*.txt or go.mod. The scripts call public registry and advisory APIs for each dependency, and the plugin README recommends a signed-in gh to avoid GitHub rate limits.

Good for

Security reviewers checking dependencies before an engagement
Maintainers who want a report on the risk in their lockfile tree
Teams comparing replacement options for abandoned or archived packages

Install supply-chain-risk-auditor

Claude Code

Copy this prompt.

Install the supply-chain-risk-auditor skill in Claude Code. Run this command:

/plugin marketplace add trailofbits/skills
/plugin install supply-chain-risk-auditor@trailofbits

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires uv.

Pick one route. The plugin holds only this skill and has no hooks or MCP config. Whichever you choose, install uv first, and sign in with gh auth login if you want the higher GitHub rate limit. There is no Claude app upload route because the skill runs Python scripts through uv.

Built from the package data we hold. Not yet tested by us. Source.

What's in the skill

When it loads

Audits a project's dependencies for supply-chain risk: version-matched advisories for direct dependencies and the full lockfile tree, abandoned or archived upstreams, npm publisher concentration, and install-time script execution. Use when asked to audit dependencies, assess supply-chain or third-party package risk, or review a dependency tree before an engagement.

Sections in SKILL.md

Why the scripts do the measuring, not you
Workflow
Style for what you add
Reading the report
Rationalizations to reject
When not to use

Key facts

Author: Trail of Bits (plugin README: Spencer Michaels original, Eric Quintero current design)
Repo: trailofbits/skills
License: CC-BY-SA-4.0 (LICENSE and README agree)
Installs on skills.sh: 8,528 (Nov 2, 2026)
GitHub stars: 7,954 for the whole repo (Nov 2, 2026)
Last repo commit: Oct 28, 2026
Audits on skills.sh: Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass
What it includes: 13 files including SKILL.md: 4 Python scripts, 4 test files, pyproject.toml, uv.lock, agents/openai.yaml and a logo
Needs: uv; optionally an authenticated gh (60 versus 5,000 GitHub requests per hour)
Part of: The supply-chain-risk-auditor plugin (1 skill, no hooks or MCP config in the plugin folder)

What it includes

Looks for package.json, pyproject.toml, requirements files or go.mod; stops if none.
Read · 1. Check manifests
Runs gh auth status and says so if GitHub access is unauthenticated.
Runs code · 2. Check gh auth
Gathers advisories, deprecations, archived repos, publisher data and install scripts.
Run with uv. SKILL.md says to put the outputs outside the audited repository unless you ask otherwise.
Try: “Audit this project's dependencies”
Runs code · 3. collect.py
Turns findings.json into a Markdown report with a coverage table.
Runs code · 3. render.py
Claude reads report.md and findings.json; the report is the deliverable.
Read · 4. Read the report
Adds a narrative, upgrade paths, replacement candidates and an install-scripts note.
SKILL.md says to label replacement ideas as judgment, not measurement.
5. Add judgment
Unassessable is not risk, and no findings is not endorsement.
The skill lists 4 rationalizations to reject, including "No findings, so the dependencies are safe."
Rules for reading results

From the official docs

Permissions

Medium risk

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

Dependency names leave your machine
collect.py sends package names and versions to OSV, npm, PyPI, the Go proxy, deps.dev, GitHub and OpenSSF Scorecard. Do not run it on a private project if those names are confidential. github.com
It reads your gh token
sources.py runs gh auth token and uses the result for api.github.com requests. Use a gh login with minimal scopes, or run unauthenticated and accept the 60 requests per hour limit. github.com
Optional pip-audit and a response cache
sources.py runs pip-audit with --no-deps and --disable-pip when it is installed. HTTP responses are cached under the system temp directory, and the plugin README says --offline reruns from that cache. github.com
It does not install or build the audited project
The plugin README says the audit reads manifests and lockfiles and never installs, builds or executes the project or its packages. github.com

Technical details

Turning on read-only	No read-only switch. The scripts write their findings and report to an output folder.
Talks to	api.osv.dev. registry.npmjs.org and api.npmjs.org. pypi.org. proxy.golang.org. api.deps.dev. api.github.com. api.scorecard.dev
Sign-in and keys	Reads the gh CLI token with gh auth token when gh is signed in. SKILL.md names no other key.
Scopes	Reads manifests and lockfiles in the project you point it at Writes findings.json and report.md to an output folder Runs uv run on two bundled Python scripts Makes HTTP requests to public registries and advisory services for each dependency

Questions

What does the supply-chain-risk-auditor skill do?

It makes Claude run a collector and a renderer over your dependencies and produce a risk report. It covers version-matched advisories, deprecated releases, archived or abandoned repos, npm publisher concentration and install-time scripts.

Which ecosystems does it support?

npm, PyPI and Go. Lockfiles it reads are package-lock.json, npm-shrinkwrap.json, uv.lock and a go 1.17+ go.mod. yarn.lock, pnpm-lock.yaml and poetry.lock are not read.

How do I install it?

Use the install box at the top of this page. The repo documents a plugin marketplace route. You also need uv. The plugin holds only this skill.

Does it need an API key?

No key is named. The plugin README recommends a signed-in gh, because unauthenticated GitHub allows 60 requests per hour and the collector makes several per dependency.

Does it run my code or install my packages?

The plugin README says the audit never installs, builds or executes the project or its packages. It reads manifests and lockfiles and queries public services.

Is it safe?

It sends your dependency names and versions to several public APIs and reads your gh token if you are signed in. Avoid it on projects whose dependency lists are confidential.

Does it work in the Claude app?

The install box has no Claude app upload route. The skill runs Python scripts through uv on your machine.

Does "no findings" mean my dependencies are safe?

No. SKILL.md says no findings means none among what was assessed, and that on PyPI and Go half the criteria cannot be assessed. Read the coverage table.

Information

updated Nov 2, 2026

Repository: trailofbits/skills
Folder: plugins/supply-chain-risk-auditor/skills/supply-chain-risk-auditor

trailofbits/skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 5

trailofbits/skills README · github.com
supply-chain-risk-auditor plugin README · github.com
trailofbits/skills marketplace.json · github.com
supply-chain-risk-auditor SKILL.md · github.com
sources.py · github.com