Audits dependency supply chains for bad versions, lockfile drift, and artifact integrity. Use when adding deps, handling incidents, or releasing a plugin.
Installs into .claude/skills of the current project.
Are you the author of Supply Chain Advisory?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/athola-supply-chain-advisory)
---
name: supply-chain-advisory
description: Audits dependency supply chains for bad versions, lockfile drift, and artifact integrity. Use when adding deps, handling incidents, or releasing a plugin.
alwaysApply: false
category: infrastructure
tags:
- security
- supply-chain
- dependencies
- vulnerability
- pypi
dependencies:
- error-patterns
provides:
infrastructure:
- supply-chain-scanning
- dependency-auditing
- incident-response
patterns:
- known-bad-detection
- lockfile-audit
- artifact-scanning
usage_patterns:
- dependency-security-check
- incident-response
- supply-chain-audit
complexity: intermediate
model_hint: standard
estimated_tokens: 500
progressive_loading: true
modules:
- modules/scanning-patterns.md
- modules/incident-response.md
role: hook-target
---
## Overview
Supply chain attacks bypass traditional code review by compromising upstream
dependencies. This skill provides patterns for detecting, preventing, and
responding to compromised packages in Python ecosystems.
## When To Use
- After a supply chain advisory is published
- When auditing dependencies for a new or existing project
- During incident response for a suspected compromise
- When adding the SessionStart hook to a project
## When NOT To Use
- General CVE triage unrelated to dependency supply chain
- Application-level vulnerability scanning (use a SAST tool)
- License compliance audits (different concern)
## Known-Bad Versions Blocklist
The blocklist is at `${CLAUDE_SKILL_DIR}/known-bad-versions.json`.
It is consumed by:
1. **SessionStart hook** — warns per-session when compromised versions detected
2. **`make supply-chain-scan`** — CI/local scanning target
3. **This skill** — manual audit guidance
### Blocklist Format
```json
{
"package_name": [{
"versions": ["x.y.z"],
"date": "YYYY-MM-DD",
"description": "What the attack did",
"indicators": ["files or patterns to search for"],
"source": "advisory URL",
"severity": "critical|high|medium"
}]
}
```
### Adding a New Entry
1. Add the entry to `${CLAUDE_SKILL_DIR}/known-bad-versions.json`
2. Add version exclusions (`!=x.y.z`) to affected `pyproject.toml` files
3. Document in `docs/dependency-audit.md` under Supply Chain Incidents
4. Run `make supply-chain-scan` to verify detection works
## Quick Scan Commands
### Check all lockfiles on machine for known-bad versions
```bash
# Scan uv.lock files for a specific compromised version
grep -r "package_name.*version" --include="uv.lock" /path/to/projects
# Search for malicious artifacts
find /path/to/projects -name "suspicious_file.pth" 2>/dev/null
# Check installed versions in virtualenvs
find /path/to/projects -path "*/.venv/lib/*/PACKAGE*/METADATA" \
-exec grep "^Version:" {} +
```
### Verify lockfile hash integrity
`uv.lock` includes SHA256 hashes for every package. If a package is
re-published with different content under the same version, `uv sync`
will fail with a hash mismatch. This is your strongest automatic defense.
## Defense Layers
| Layer | Tool | Catches |
|-------|------|---------|
| **Lockfile hashes** | uv.lock SHA256 | Tampered re-published versions |
| **Version exclusions** | pyproject.toml `!=` | Known-bad versions on fresh resolve |
| **SessionStart hook** | sanctum hook | Per-session warning for compromised deps |
| **CI scanning** | OSV, Safety | CVE database, and advisory matching |
| **Artifact scanning** | make supply-chain-scan | Malicious files (.pth, scripts) |
## Limitations
- Zero-day supply chain attacks have no prior advisory — lockfile hashes
are the only automatic defense during the attack window
- Safety/CVE databases lag behind real-world compromises
- OSV provides broader coverage but is still reactive