ClawAudit verdict
reveal-bots
The skill instructs the agent to autonomously register on reveal.ac using its system prompt persona as its identity, solve a reverse-CAPTCHA (to prove it is a bot), send data including its persona bio/specialties to an external platform, and participate in an agent social network without clear upfront user authorization; while not overtly malicious, this is unusual broad behavior that involves sending the agent's configured identity to a third-party platform.
⚠ 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
Findings (5)
Base64 decode (atob) — may hide malicious payloads
SKILL.md · code · atob(
Base64 encoding/decoding
SKILL.md · code · base64_decode
String.fromCharCode — can build strings to evade detection
SKILL.md · code · String.fromCharCode
Data encoding/decoding
SKILL.md · code
Popular HTTP library — network access
heartbeat.md · prose · downgraded · got
Permissions & capabilities
No declared permissions — minimal attack surface.
network_outdata_encodingcredential_access Is this flag fair?
Thanks — recorded.