Skip to content
Back to skills

Levn Ln 762 Dependency Audit

ASecurity

Audits project dependencies for vulnerabilities. Multi-ecosystem support (npm, .NET, Python, Go). CVSS-based severity classification.

  • 666 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added February 9, 2026
securitypythonrustgorubyphpci/cdsecurity

Security analysis

A100/100

Scanned February 12, 2026

npx -y skills add majiayu000/claude-skill-registry --skill levn-ln-762-dependency-audit --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Levn Ln 762 Dependency Audit?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Levn Ln 762 Dependency Audit
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/majiayu000-levn-ln-762-dependency-audit/badge)](https://www.skillsdirectory.com/skills/majiayu000-levn-ln-762-dependency-audit)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: ln-762-dependency-audit
description: Audits project dependencies for vulnerabilities. Multi-ecosystem support (npm, .NET, Python, Go). CVSS-based severity classification.
---

# Dependency Audit

Audits project dependencies for known security vulnerabilities across multiple package ecosystems.

## Purpose & Scope

- Detect vulnerable dependencies using ecosystem-specific tools
- Support multiple ecosystems: npm, NuGet, pip, Go modules, Bundler, Cargo
- Classify vulnerabilities by severity (Critical/High/Medium/Low)
- Provide fix recommendations with safe auto-fix guidance
- Return normalized report to parent orchestrator (ln-760)

## When to Use

- During project bootstrap (via ln-760-security-setup)
- CI/CD pipeline security checks
- Pre-release security validation
- Regular scheduled audits

---

## Workflow

### Phase 1: Ecosystem Detection

**Step 1: Detect Package Managers**
- Check for `package.json` / `package-lock.json` (npm)
- Check for `*.csproj` / `packages.config` (.NET)
- Check for `requirements.txt` / `Pipfile` / `pyproject.toml` (Python)
- Check for `go.mod` (Go)
- Check for `Gemfile` (Ruby), `Cargo.toml` (Rust), `composer.json` (PHP)

**Step 2: Check Tool Availability**
- For each detected ecosystem, verify audit tool is available
- If tool missing: log warning, skip ecosystem (do not fail)

### Phase 2: Audit Execution

**Step 1: Run Ecosystem Audits**
- Execute audit command for each detected ecosystem
- Prefer JSON output for parsing (see `references/audit_commands.md`)
- Run audits in parallel where possible

**Step 2: Parse Results**
- Normalize findings to common format: package, version, vulnerability ID, severity
- Extract CVSS score if available

### Phase 3: Report Generation

**Step 1: Severity Classification**
- Map CVSS scores to severity per `references/severity_mapping.md`
- Critical: CVSS 9.0-10.0
- High: CVSS 7.0-8.9
- Medium: CVSS 4.0-6.9
- Low: CVSS 0.1-3.9

**Step 2: Group and Sort**
- Group by ecosystem
- Sort by severity (Critical first)
- Include vulnerability count summary

**Step 3: Build Report**
- Include package name, current version, fixed version
- Include vulnerability ID (CVE/GHSA/OSV)
- Do NOT include exploit details

### Phase 4: Fix Recommendations

**Step 1: Classify Fix Type**
- Patch update (safe auto-fix)
- Minor update (usually safe)
- Major update (manual review required)
- No fix available (document and monitor)

**Step 2: Generate Recommendations**
- For each vulnerability: suggest fix command
- Flag breaking changes if major version bump
- Note if fix requires code changes

**Step 3: Return Results**
- Return structured report to orchestrator
- Include summary: packages audited, vulnerabilities found, by severity

---

## Critical Rules

1. **Never auto-fix major versions** - may introduce breaking changes
2. **Verify lock file integrity** - regenerate if corrupted
3. **Respect severity thresholds** - per environment (see `references/severity_mapping.md`)
4. **Document unfixable vulns** - add to known issues with review date
5. **No exploit code** - report IDs only, not exploitation details

---

## Definition of Done

- [ ] All detected ecosystems audited
- [ ] Findings classified by severity with CVSS mapping
- [ ] Fix recommendations provided (safe vs manual)
- [ ] Report in normalized format returned
- [ ] Critical vulnerabilities prominently flagged
- [ ] Lock file integrity verified

---

## Reference Files

| File | Purpose |
|------|---------|
| `references/audit_commands.md` | Ecosystem-specific audit commands |
| `references/severity_mapping.md` | CVSS to severity level mapping |
| `references/ci_integration_guide.md` | CI/CD integration guidance |

---

**Version:** 2.0.0
**Last Updated:** 2026-01-10

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.