ClawAudit verdict
agi-farm
This skill is a multi-agent team setup wizard that writes local workspace files, installs packages, and optionally pushes to GitHub — all transparent operations consistent with its stated purpose of bootstrapping an OpenClaw agent team.
⚠ Flagged for review — coarse, uncorroborated signal, not a confirmed exploit. Review the config yourself before installing.
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.
Permission integrity
network_out
file_read+write
package_install
Findings (10)
Pipe to python — executes piped content as Python code
SKILL.md · code · | python3
Pipe-to-python pattern — remote code execution risk
SKILL.md · code · curl -s http://localhost:8080/api/data | python
<script> tag in markdown — potential code injection
dashboard.py · prose · downgraded · <script>
Blob URL — may embed executable content
SKILL.md · code · blob:
References agent memory files
references/dashboard.md · code · MEMORY.md
pip3 install — installs Python packages at runtime
references/dashboard.md · code · pip3 install
subprocess execution — runs system commands from Python
dashboard.py · prose · downgraded · subprocess.run(
References agent configuration files
README.md · code · CLAUDE.md
Python threading.Timer — deferred execution
dashboard.py · prose · downgraded · threading.Timer(
Python os.environ.get — reads environment variable
scripts/auto-dispatch.py · prose · downgraded · os.environ.get(
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_innetwork_outpackage_installfile_write Is this flag fair?
Thanks — recorded.