workbench-capy-process-check
Pass
Audited by Gen Agent Trust Hub on May 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were detected. The skill is designed with defensive instructions to prevent common agent security risks.\n- [DATA_EXFILTRATION]: The instructions explicitly forbid the exfiltration of sensitive session data.\n
- Evidence: Rule states not to copy private thread URLs, screenshots, raw logs, cookies, OAuth codes, or tokens into public docs.\n- [COMMAND_EXECUTION]: The skill implements a manual checkpoint for high-risk operations.\n
- Evidence: Rules prohibit clicking OAuth, merge, publish, or destructive controls without explicit human approval for the exact action.\n- [CREDENTIALS_UNSAFE]: The skill contains a fail-safe for unexpected authentication requests.\n
- Evidence: Instructions mandate stopping the task if the browser page asks for credentials or new access unless previously approved.
Audit Metadata