diagnose-hard-problem
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data such as network traces, log dumps, and user-provided error messages to diagnose issues.\n
- Ingestion points: Data enters through captured artifacts in Phase 1 and 2, and human input via the
scripts/hitl-loop.template.shscript.\n - Boundary markers: No specific machine-readable boundary markers are defined for data isolation, though Markdown structure is used.\n
- Capability inventory: The agent executes shell commands, CLI tools, and test harnesses.\n
- Sanitization: The skill contains explicit, mandatory instructions to redact secrets and credentials before processing or displaying them, and specifically advises that sensitive steps like signing in be performed by the user without agent observation.\n- [COMMAND_EXECUTION]: The skill requires running CLI commands,
curl, and test runners to establish feedback loops for bug reproduction.\n- [DYNAMIC_EXECUTION]: The skill methodology involves generating and running temporary test harnesses, browser scripts, and human-in-the-loop scripts to verify bug fixes.
Audit Metadata