agentleFS
Lethe044Lethe044/agentskill-kit/skills/requirements-auditor/SKILL.md

requirements-auditor

Skill50 starsChanged 21 days agoMITView on GitHub

Audits a Python requirements.txt for unpinned packages, open version ranges, duplicate entries, installs from URLs or VCS and extra package indexes that invite dependency confusion. Use when the user asks to review dependencies, make builds reproducible or check a requirements file before release.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

Requirements auditor

Checks the shape of a requirements.txt. It does not look up vulnerabilities, so pair it with a vulnerability scanner when the user needs one.

When to use

  • The user asks whether dependencies are pinned or builds are reproducible.
  • A requirements file grew over time and may contain duplicates.
  • A project installs from a private index and the user wants to know the risks.

How to run

python scripts/audit_requirements.py requirements.txt

Rules

RuleLevelMeaning
RQ001warningpackage has no version specifier
RQ002infoonly a lower bound (>=) or a wildcard, the upper bound is open
RQ003warningpackage is listed twice (names are compared after normalizing -, _ and .)
RQ004warninginstalled from a URL or VCS without a pinned commit or tag
RQ005warning--extra-index-url lets a public package replace a private one with the same name
RQ006warning--trusted-host disables TLS verification for a host

Lines that start with -r or -c, comments, blank lines, extras (pkg[extra]) and environment markers (; python_version < "3.12") are handled.

Output

JSON with requirements (count) and findings, each {"line", "rule", "severity", "name", "message"}. The exit code is 1 when there is at least one warning.

After running

Suggest pip-compile or uv pip compile to produce a fully pinned lock file with hashes, keep the loose file as the input, and use --index-url with a single trusted index when possible.

Errors

A missing file exits with code 2.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.