ClawAudit verdict

skill-vetting

88
🟢 Trusted
Low risk — reviewed by ClawAudit, behavior matches stated purpose

Fetches from network AND uses eval/Function

This skill provides a structured workflow to vet ClawHub skills by downloading to /tmp (not workspace), running an automated scanner, and performing manual review with explicit prompt injection warnings; notably it warns reviewers not to follow instructions found in skill files, which is responsible security tooling.

Automated static analysis — not a human review. ClawAudit flags capabilities, not confirmed intent, and can produce false positives. Disagree with this verdict? Use Dispute below.

0
security
70
transparency
70
maintenance

What it does

These are capability combinations: each listed behavior occurs in the skill, but ClawAudit detects co-occurrence — it does not verify that one flows into another. Read the code to confirm a live chain.

Capability combination critical

Fetches from network AND uses eval/Function — the remote-code-execution pattern (data-flow not verified)

LLM05 · LLM06 · ASI05

Capability combination critical

Encodes data AND uses eval — the obfuscated-execution pattern (atob + eval; data-flow not verified)

LLM05 · ASI05 · ASI10

Permission integrity

Makes network requests but does not declare curl/wget in required binaries

network_out

Findings (17)

Pattern match critical

Uses eval() — can execute arbitrary code

ARCHITECTURE.md · code · eval(

Pattern match critical

Dynamic __import__('os') — Python OS command execution

references/patterns.md · code · __import__('os')

Pattern match critical

Recursive delete from root or home — destructive command

references/patterns.md · code · rm -rf /

Pattern match critical

Accesses sensitive system files

references/patterns.md · code · /etc/passwd

Pattern match critical

Prompt injection — tries to override agent instructions

references/patterns.md · code · Ignore all previous instructions

Pattern match high

Uses exec() — may execute shell commands

references/patterns.md · code · exec(

Pattern match high

Hex-encoded string — possible obfuscated payload

references/patterns.md · code · \x69\x6d\x70\x6f\x72\x74\x20\x6f\x73

Pattern match high

Accesses .ssh directory

references/patterns.md · code · .ssh/

Pattern match high

Possible prompt injection — attempts to redefine agent identity

references/patterns.md · code · You are now

Pattern match high

Fake system prompt — attempts to inject instructions

references/patterns.md · code · system: You are

Pattern match high

subprocess execution — runs system commands from Python

references/patterns.md · code · subprocess.run(

Pattern match high

subprocess with shell=True — command injection vector

references/patterns.md · code · subprocess.run(f'ls {user_input}', shell=True

Pattern match high

os.system/popen — direct OS command execution

references/patterns.md · code · os.system(

Pattern match medium

importlib.import_module — dynamic module loading

SKILL.md · code · importlib.import_module(

Pattern match medium

Python shutil file operation — copies/moves/deletes files

references/patterns.md · code · shutil.rmtree(

Pattern match medium

POSTs data to external URL

references/patterns.md · code · .post('https://

Pattern match medium

Python os.getenv — reads environment variable

references/patterns.md · code · os.getenv(

Why the tier is capped

Execution sink present in raw bytes (Hard Floor: class D). Final tier capped at Caution — cannot be lifted by any downgrade, example-payload opt-in, or allowlist.

Permissions & capabilities

No declared permissions — minimal attack surface.

network_outdynamic_evaldata_encoding
Check another skill Browse the registry Auditing your own skills or configs? Use the API