Cpython Security Toolkit — Security skill for Claude Code
A Claude Code plugin for finding semantic security flaws in CPython's standard library — archive extraction boundaries, validation coverage gaps, resource amplification, and audit hook bypass.
How to install Cpython Security Toolkit
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open ReviewToolkits/cpython-security-toolkit and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Cpython Security Toolkit does
A Claude Code plugin for finding semantic security flaws in CPython's standard library — archive extraction boundaries, validation coverage gaps, resource amplification, and audit hook bypass.
Alternatives in Security
- Security Threat Model — Generate repo-specific threat models identifying trust boundaries 14.6k ★
- Security Guidance — Hook that warns about command injection, XSS, and unsafe patterns 14k ★
- Security Guardian — Comprehensive security analysis for RTK CLI tool, focusing on command injection, shell escaping, hook security 11.9k ★
README
CPython Security Toolkit
A [Claude Code](https://docs.anthropic.com/en/docs/claude-code) plugin for finding semantic security flaws in CPython's standard library — the class of bug that generic SAST tools, fuzzers, and sanitizers do not model well.
Built around CPython-specific security invariants: archive extraction boundaries, protocol validation coverage, resource amplification ratios, and audit hook gaps. Every detector is anchored to a confirmed historical CVE; every finding requires a runnable reproducer before it reaches a maintainer.
⚠️ Read this before you use this toolkit on CPython
This tool finds **security vulnerabilities** in CPython's standard library. Security findings are handled differently from correctness bugs.
- Read WORKING_WITH_MAINTAINERS.md. It is the most important document in this repository.
- Security findings go to
security@python.orgfirst — never the public issue tracker. See the CPython security policy. - Pre-triage with a trusted CPython developer before submitting. One private message — "I found something that might be a security issue in
tarfile, can you take a look?" — takes five minutes and saves everyone time. - Never file a finding without a runnable reproducer. A finding with no reproducer is not a finding. The tool enforces this.
- Concise and objective reports only. CPython maintainers are busy. One clear sentence describing the violated invariant is worth more than three paragraphs of analysis.
Why a Separate Tool?
| Concern | cpython-review-toolkit | cpython-security-toolkit |
|---|---|---|
| Layer | C source (Modules/, Objects/, Python/) | Python stdlib (Lib/) |
| Question asked | "Is this C implementation correct?" | "Does this stdlib behavior violate a security invariant?" |
| Bug class | Refcount leaks, null safety, error paths | Archive escapes, validation bypass, resource ampl |
Related Skills
Hackz Huntkit
Offensive-security playbooks for AI coding agents — IDOR, 403 bypass, CORS, finding validation and dupe-checki
Audit Claude Components
Comprehensive audit of .claude folder. Checks structure, quality, cross-references, antipatterns, resource usa
Sleep
Standard end-of-session consolidation — FAST. Reflection, state banner, journal, archive, commit. The full aud
Deep Sleep
DEEP consolidation — full memory audit, scouts, belief evidence, template sync, summary backfill. Run before/a
Claudemd Audit
Aggregate claudemd rule-hits over the last N days. Shows top banned patterns, hook deny counts, per-spec-secti
Askit Evaluate
Evaluate a skill or plugin against the Advanced Skill Library Standard and report per-rule findings, the tier,
Related Agents
Auth Tester
Authentication and session management testing agent. Use for login bypass, session fixation, password reset fl
Bizlogic Hunter
Delegates to this agent when the user wants to test for business logic flaws, find workflow bypass vulnerabili
Deep Review Guard Bypass
Traces guard functions (permission, capacity, rate-limit, validation) to find any code path that reaches a pro