How do I install the Skill Security Auditor V2 skill?

Run npx skillmds@latest add nickgallick/skill-security-auditor-v2 in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Skill Security Auditor V2 safe to use?

SkillMD's automated safety review verdict for this skill is PASS. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: executes scripts, makes network calls, reads secrets. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Skill Security Auditor V2?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Skill Security Auditor V2 free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Skill Security Auditor V2?

nickgallick (@nickgallick) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Skill Security Auditor V2 by nickgallick

Name: Skill Security Auditor V2
Rating: 4.666666666666667 (3 reviews)
Author: nickgallick

Skill Security Auditor V2

Hybrid security auditor for OpenClaw skills, Claude/Codex skills, and app repos. Use when installing a new skill, auditing a repo before use or deploy, reviewing custom scripts, checking for prompt injection, command execution, data exfiltration, dependency risk, secrets exposure, or privilege escalation. Use as the default gatekeeper before installing any third-party skill.

nickgallick Updated Aug 19, 2026 0 repo stars

File contents

Skill Security Auditor v2

Run a hybrid security audit before installing a skill or deploying code. Return a clear PASS / WARN / FAIL verdict with concrete findings and remediation.

Use

Audit a local path

python3 scripts/security_auditor_v2.py /path/to/target

Audit a git repo

python3 scripts/security_auditor_v2.py https://github.com/org/repo

JSON output

python3 scripts/security_auditor_v2.py /path/to/target --json

Strict mode

python3 scripts/security_auditor_v2.py /path/to/target --strict

What it checks

Command execution risk
- shell execution
- eval/exec
- dynamic imports
- child_process / subprocess misuse
Prompt injection risk
- instruction override attempts
- role hijacking
- safety bypass language
- hidden directives
Data exfiltration risk
- outbound network writes
- credential harvesting patterns
- suspicious file access
Privilege and persistence risk
- sudo
- cron modification
- shell/profile tampering
- dangerous permissions
Filesystem safety
- writes outside expected scope
- destructive deletes
- symlinks
- hidden sensitive files
Dependency and secret risk
- suspicious install-at-runtime behavior
- unpinned deps
- likely secrets/tokens in repo
Optional deep scanners if installed
- Semgrep
- Bandit
- Gitleaks
- Trivy

Output contract

Always return:

PASS: no meaningful issues found
WARN: review required, but likely safe after inspection
FAIL: unsafe to install/use as-is

For each finding include:

severity
category
file and line when possible
why it matters
how to fix it

Hard rules

Treat third-party skills as untrusted until scanned
Never auto-approve a FAIL result
For WARN, explain why it is likely safe or why manual review is needed
Distinguish live executable risk from documentation/examples
Prefer fewer high-confidence findings over noisy output

Optional deeper tools

If available on the machine, the script will automatically use:

semgrep
bandit
gitleaks
trivy

If missing, the script still works with the native scanner only.

References

Read references/risk-model.md when tuning findings or reviewing gray-area cases.

---name: skill-security-auditor-v2description: Hybrid security auditor for OpenClaw skills, Claude/Codex skills, and app repos. Use when installing a new skill, auditing a repo before use or deploy, reviewing custom scripts, checking for prompt injection, command execution, data exfiltration, dependency risk, secrets exposure, or privilege escalation. Use as the default gatekeeper before installing any third-party skill.---# Skill Security Auditor v2Run a hybrid security audit before installing a skill or deploying code. Return a clear **PASS / WARN / FAIL** verdict with concrete findings and remediation.## Use### Audit a local path```bashpython3 scripts/security_auditor_v2.py /path/to/target```### Audit a git repo```bashpython3 scripts/security_auditor_v2.py https://github.com/org/repo```### JSON output```bashpython3 scripts/security_auditor_v2.py /path/to/target --json```### Strict mode```bashpython3 scripts/security_auditor_v2.py /path/to/target --strict```## What it checks1. **Command execution risk** - shell execution - eval/exec - dynamic imports - `child_process` / subprocess misuse2. **Prompt injection risk** - instruction override attempts - role hijacking - safety bypass language - hidden directives3. **Data exfiltration risk** - outbound network writes - credential harvesting patterns - suspicious file access4. **Privilege and persistence risk** - `sudo` - cron modification - shell/profile tampering - dangerous permissions5. **Filesystem safety** - writes outside expected scope - destructive deletes - symlinks - hidden sensitive files6. **Dependency and secret risk** - suspicious install-at-runtime behavior - unpinned deps - likely secrets/tokens in repo7. **Optional deep scanners if installed** - Semgrep - Bandit - Gitleaks - Trivy## Output contractAlways return:- **PASS**: no meaningful issues found- **WARN**: review required, but likely safe after inspection- **FAIL**: unsafe to install/use as-isFor each finding include:- severity- category- file and line when possible- why it matters- how to fix it## Hard rules- Treat third-party skills as untrusted until scanned- Never auto-approve a FAIL result- For WARN, explain why it is likely safe or why manual review is needed- Distinguish **live executable risk** from **documentation/examples**- Prefer fewer high-confidence findings over noisy output## Optional deeper toolsIf available on the machine, the script will automatically use:- `semgrep`- `bandit`- `gitleaks`- `trivy`If missing, the script still works with the native scanner only.## References- Read `references/risk-model.md` when tuning findings or reviewing gray-area cases.

Frequently asked questions