Skip to main content
skill CX Codex

dependency-auditor · useai-pro — Claude Code skill: what it does, how to install

Audits npm, pip, and Go packages before installation. Protects projects from typosquatting, dangerous install scripts, and vulnerable transitive dependencies.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add UseAI-pro/openclaw-skills-security --skill dependency-auditor -a claude-code
Open on GitHub

What dependency-auditor does

This skill evaluates packages whenever another agent suggests adding new dependencies to your project. It inspects package names against common typosquat patterns, tracks publication recency, and flags oversized dependency trees exceeding 100 transitive components. It also inspects package.json configurations to flag lifecycle scripts like preinstall or postinstall before they execute with system privileges.

The skill operates in an offline posture because its manifest explicitly disables network calls (network: false). It does not query remote registries on its own; instead, it guides the agent to interpret local audit tools such as npm audit, pip-audit, or govulncheck. Every audit ends with a formatted report marking libraries as APPROVE, REVIEW, or REJECT alongside actionable remediation advice.

What it helps with

  • Detect typosquatted package names before running package managers
  • Uncover suspicious preinstall and postinstall scripts in dependencies
  • Filter out risky copyleft licenses like GPL-3.0 or AGPL-3.0
  • Audit deep transitive trees and CVSS severity using local scanners

Example prompts for the agent

  • Audit these packages before running npm install
  • Check requirements.txt for typosquatting and suspicious maintainers
  • Review security risks for the Go module proposed by the agent

How to install dependency-auditor

1Skills CLI (Vercel)

npx skills add UseAI-pro/openclaw-skills-security --skill dependency-auditor -a claude-code

2Manually: put the files in place

~/.claude/skills/dependency-auditor/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/dependency-auditor/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about dependency-auditor

How to install dependency-auditor in Claude Code?

Run npx skills add UseAI-pro/openclaw-skills-security --skill dependency-auditor -a claude-code in your terminal. You can also place the configuration file directly at ~/.claude/skills/dependency-auditor/SKILL.md.

What does dependency-auditor inspect?

It audits naming authenticity, suspicious install hooks, transitive dependency counts over 100, and software licenses like MIT or AGPL.

Does dependency-auditor make external network requests?

No, the skill runs with network: false permissions. It relies on local scanners like npm audit, pip-audit, or govulncheck to inspect vulnerabilities.

Source text

SKILL.md · MIT
---
name: dependency-auditor
description: Audit npm, pip, and Go dependencies that OpenClaw skills try to install. Checks for known vulnerabilities, typosquatting,
  and malicious packages.
metadata:
  short-description: Audit skill-installed dependencies for typosquatting, hooks, and known package risk.
  why: Catch supply-chain abuse before a skill pulls malicious packages into the environment.
  what: Provides a dependency review module for npm, pip, and Go install flows used by OpenClaw skills.
  how: Checks package identity, install hooks, recency, reputation, and vulnerability severity.
  results: Produces dependency findings with install recommendations and block conditions.
  version: 1.0.0
  updated: '2026-03-10T03:42:30Z'
  jtbd-1: When a skill wants to install packages and I need a quick supply-chain risk review first.
  audit:
    kind: module
    author: useclawpro
    category: Security
    trust-score: 93
    last-audited: '2026-02-03'
    permissions:
      file-read: true
      file-write: false
      network: false
      shell: false
---

# Dependency Auditor

You are a dependency security auditor for OpenClaw. When a skill tries to install packages or you review a project's dependencies, check for security issues.

## When to Audit

- Before running `npm install`, `pip install`, `go get` commands suggested by a skill
- When reviewing a skill that adds dependencies to package.json or requirements.txt
- When a skill suggests installing a package you haven't used before
- During periodic security audits of your project

## Audit Checklist

### 1. Package Legitimacy

For each package, verify:

- [ ] **Name matches intent** — is it the actual package, or a typosquat?
  ```
  lodash     ← legitimate
  l0dash     ← typosquat (zero instead of 'o')
  lodash-es  ← legitimate variant
  lodash-ess ← typosquat (extra 's')
  ```

- [ ] **Publisher is known** — check who published the package
  ```
  npm: Check npmjs.com/package/<name> for publisher identity
  pip: Check pypi.org/project/<name> for maintainer
  ```

- [ ] **Download count is reasonable** — very new packages with 0-10 downloads are higher risk

- [ ] **Repository exists** — the package should link to a real source repository

- [ ] **Last published recently** — abandoned packages may have known unpatched vulnerabilities

### 2. Known Vulnerabilities

Check against vulnerability databases.

Note (offline-first): this skill declares `network: false`, so you must not fetch live URLs yourself. Treat links below as **manual references** for the user to open, and prefer local commands (`npm audit`, `pip-audit`, `govulncheck`) when possible.

```
NPM:
  npm audit
  Check: https://github.com/advisories

PyPI:
  pip-audit
  Check: https://osv.dev

Go:
  govulncheck
  Check: https://vuln.go.dev
```

**Severity classification:**
| Severity | Action |
|---|---|
| Critical (CVSS 9.0+) | Do not install. Find alternative. |
| High (CVSS 7.0-8.9) | Install only if patched version available. |
| Medium (CVSS 4.0-6.9) | Install with awareness. Monitor for patches. |
| Low (CVSS 0.1-3.9) | Generally acceptable. Note for future. |

### 3. Suspicious Package Indicators

**Red flags that warrant deeper investigation:**

- Package has `postinstall`, `preinstall`, or `install` scripts
  ```json
  // package.json — check "scripts" section
  "scripts": {
    "postinstall": "node setup.js"  // ← What does this do?
  }
  ```

- Package imports `child_process`, `net`, `dns`, `http` in unexpected ways

- Package reads environment variables or file system on import

- Package has obfuscated or minified source code (unusual for npm packages)

- Package was published very recently (< 1 week) and has minimal downloads

- Package name is similar to a popular package but from a different publisher

- Package has been transferred to a new owner recently

### 4. Dependency Tree Depth

Check transitive dependencies:

```
Direct dependency → sub-dependency → sub-sub-dependency
     (you audit)      (who audits?)     (nobody audits?)
```

- Flag packages with excessive dependency trees (100+ transitive deps)
- Check if any transitive dependency has known vulnerabilities
- Prefer packages with fewer dependencies

### 5. License Compatibility

Verify licenses are compatible with your project:

| License | Commercial Use | Copyleft Risk |
|---|---|---|
| MIT, ISC, BSD | Yes | No |
| Apache-2.0 | Yes | No |
| GPL-3.0 | Caution | Yes — derivative works must be GPL |
| AGPL-3.0 | Caution | Yes — even network use triggers copyleft |
| UNLICENSED | No | Unknown — avoid |

## Output Format

```
DEPENDENCY AUDIT REPORT
=======================
Package: <name>@<version>
Registry: npm / pypi / go
Requested by: <skill name or user>

CHECKS:
  [PASS] Name verification — no typosquatting detected
  [PASS] Publisher — @official-org, verified
  [WARN] Vulnerabilities — 1 medium severity (CVE-2026-XXXXX)
  [PASS] Install scripts — none
  [PASS] License — MIT
  [WARN] Dependencies — 47 transitive dependencies

OVERALL: APPROVE / REVIEW / REJECT

RECOMMENDATIONS:
  - Update to version X.Y.Z to resolve CVE-2026-XXXXX
  - Consider alternative package 'safer-alternative' with fewer dependencies
```

## Common Typosquatting Patterns

Watch for these naming tricks:

| Technique | Legitimate | Typosquat |
|---|---|---|
| Character swap | express | exrpess |
| Missing character | request | requst |
| Extra character | lodash | lodashs |
| Homoglyph | babel | babe1 (L → 1) |
| Scope confusion | @types/node | @tyeps/node |
| Hyphen trick | react-dom | react_dom |
| Prefix/suffix | webpack | webpack-tool |

## Rules

1. Never auto-approve `npm install` or `pip install` from untrusted skills
2. Always check install scripts before running — they execute with full system access
3. Pin dependency versions in production — avoid `^` or `~` ranges for security-critical packages
4. If a skill wants to install 10+ packages, review each one individually
5. When in doubt, read the package source code — it's usually small enough to skim

© UseAI-pro · MIT · UseAI-pro/openclaw-skills-security

Similar skills

All skills for Security →

dependency-auditor by other authors comes first.

dependency-auditor Alireza Rezvani An offline Claude Code skill that checks dependencies for vulnerability patterns and license conflicts, then drafts a risk-ranked upgrade plan. Use it before a release or license review. 688installs dependency-auditor OneWave-AI Checks npm packages for vulnerabilities, outdated versions and unused entries. Useful before a release or when package.json has grown messy. CX Codex +1 302installs git-guardrails-claude-code Matt Pocock This skill adds a Claude Code hook that stops risky Git commands before they run. Use it when an agent must not push or discard working-tree changes. 499,069installs ponytail-review DietrichGebert ponytail-review reviews a change beyond the diff, checking connected code, bugs, risks, load, tests, speed, and unnecessary code. Use it for PRs and commits. CX Codex +1 57,000installs ponytail-audit DietrichGebert It audits a repository and ranks bugs, risks, and redundant code in a one-shot report. Use it when you need a review without code changes. CX Codex +1 45,400installs security-audit cloudflare Reviews codebases and APIs for vulnerabilities using safe sandboxing. Fits pre-release security reviews, pen-test triage, and architecture checks. CX Codex +1 54,800installs security-and-hardening addyosmani Guides threat modeling and security checks for code that handles user input, authentication, sessions, or external services. Use it when building a feature that accepts untrusted data or reviewing code before release. CX Codex +1 70,200installs supabase official supabase Guides the agent through database configuration and strict RLS policies in Supabase. CX Codex GP ChatGPT 8installs

From the same set12

skill-vetter Reviews OpenClaw SKILL.md files before installation. Checks for typosquatting, credential file access, and dangerous permission combos like network plus shell. CX Codex +1 21,100installs skill-guard Watches active OpenClaw skills: which files they touch, where they connect, what they run in the shell. Useful when you install someone else's skill and want to see it stay inside its declared permissions. CX Codex +1 573installs prompt-guard This skill scans skill files, user messages and external data for attempts to hijack an agent's instructions. Useful before you install someone else's skill or feed the agent web content. CX Codex +1 546installs credential-scanner Scans your project for exposed keys and passwords before a file-reading skill can pick them up. CX Codex +1 538installs sandbox-guard Builds Docker sandbox configs so you can run untrusted OpenClaw skills without handing them the host. Useful the moment you install someone else's skill. CX Codex +1 483installs permission-auditor Walks through the permissions an OpenClaw skill asks for, explains each one in plain language, and suggests a smaller set. CX Codex +1 476installs output-sanitizer A skill that scans the agent's reply before it is shown or logged and strips leaked keys, PII and internal paths. Useful when output leaves your machine. CX Codex +1 475installs setup-auditor Checks an OpenClaw environment for leaked credentials, unsafe defaults and missing sandbox controls. Returns a report with a readiness verdict and a fix checklist. CX Codex +1 469installs

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor