ClawAudit verdict
skill-to-agent
skill-to-agent-converter
A workflow skill that converts OpenClaw skills into agents by creating configuration files and directories in the local workspace; process_exec is used for standard agent management commands matching the stated purpose.
⚠ 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.
Findings (9)
Recursive delete from root or home — destructive command
references/troubleshooting.md · code · rm -rf ~
Uses eval() — can execute arbitrary code
references/war-room-example.md · code · eval(
References sudo — requests elevated privileges
references/troubleshooting.md · code · sudo
Accesses OpenClaw config/secrets directly
references/troubleshooting.md · code · ~/.openclaw/openclaw.json
Possible prompt injection — attempts to redefine agent identity
references/war-room-example.md · code · You are now
References child_process — can spawn system processes
references/war-room-example.md · code · child_process
References agent memory files
references/troubleshooting.md · code · MEMORY.md
Sets world-executable permissions
references/troubleshooting.md · code · chmod 755
Popular HTTP library — network access
references/war-room-example.md · prose · downgraded · got
Why the tier is capped
Execution sink present in raw bytes (Hard Floor: class B). 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.
process_exec Is this flag fair?
Thanks — recorded.