skill-vetter-plus

Security scanner for AI agent skills. 9 built-in detection signatures. Identifies secrets, unsafe execution patterns, and prompt injection. Sub-50ms results.

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "skill-vetter-plus" with this command: npx skills add certainlogicai/skill-vetter-plus

Skill Vetter Plus

What It Does

Scans AI agent skills for security issues:

  • 9 built-in detection signatures (secrets, execution, prompt injection)
  • Sub-50ms scan time
  • Run before installing any unknown skill

How to Use

# Scan any skill directory
python3 scripts/vetter.py /path/to/installed/skill

# JSON output for piping
python3 scripts/vetter.py /path/to/skill --json

Signatures

IDSeverityWhat It Finds
hardcoded-api-keyhighapi_key, api-key
hardcoded-secrethighsecret_key, secret-token, auth_token
hardcoded-passwordhighpassword
unsafe-evalcriticaleval(
unsafe-execcriticalexec(
unsafe-os-systemcriticalos.system(
subprocess-shell-truehighshell=True
raw-networkmediumurllib.request, requests.post/get
prompt-injectioncriticalignore previous instructions, ignore the above

What It Does NOT Do

  • No AST analysis (text matching only)
  • Cannot detect control-flow obfuscation
  • Cannot analyze compiled binaries
  • Not a replacement for manual code review

Results

Scanned 12 files in 23ms
Found 1 issue(s):
  [CRITICAL] unsafe-eval at /skill/scripts/mail.py:45
    → eval() can execute arbitrary code (matched: 'eval(')

Pro Upgrade

FeaturePro ($49)
Real-time scanning
Weekly signature updates
Team sharing
Custom signatures
Priority support

Limitations

  • Text-based pattern matching
  • Cannot detect all malware — only patterns in the signature database

Attribution

  • Built by: CertainLogic
  • Concept: Skill security pre-checking (industry standard practice)

Links


Built by CertainLogic

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Security

memory-encrypted

本地加密存储记忆,定时自动备份到指定目录。防止记忆泄露,适合企业付费用户。用于用户提到加密、备份、安全、私密等场景。

Registry SourceRecently Updated
Security

Ultimate Skills Finder

The ultimate multi-source OpenClaw/agent skills finder. Searches across ClawHub, skills.sh, Rush registry, LobeHub, SkillsMP, llmbase.ai, skillsllm.com, and...

Registry SourceRecently Updated
Security

Flue

Let agents control many desktop software directly from the cli, with one pip install, and no MCP servers.

Registry SourceRecently Updated
1481Profile unavailable
Security

IronClaw Security Guard

Add lightweight defense-in-depth guardrails to OpenClaw with dangerous-command blocking, prompt-injection detection, secret redaction, and audit logging.

Registry SourceRecently Updated
1420Profile unavailable