diagnose
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data such as bug reports and error logs.
- Ingestion points: Bug reports, error outputs, and user symptoms are ingested during feedback loop construction (SKILL.md, Step 1).
- Boundary markers: The skill uses a structured 'Lean Delivery format' to present findings and stage changes (SKILL.md, Step 3).
- Capability inventory: The skill can execute shell commands, run test harnesses, and modify codebase files (SKILL.md, Execution Protocol).
- Sanitization: A mandatory 'Redaction Guardrail' is defined to replace secrets, tokens, and API keys with placeholders in all outputs and logs (SKILL.md, Operating Invariants).
- [DYNAMIC_EXECUTION]: The skill dynamically generates and executes reproduction scripts and diagnostic harnesses.
- Evidence: Temporary diagnostic scripts and harnesses are generated and executed as part of the feedback loop hierarchy (SKILL.md, Section 2).
- Control Measures: All temporary artifacts are strictly confined to the
.agents/scratch/directory and must be deleted upon completion (SKILL.md, Operating Invariants). - [COMMAND_EXECUTION]: The skill executes system commands to verify regressions and fixes.
- Evidence: Feedback loops include direct CLI invocations and HTTP curl commands (SKILL.md, Section 2).
- Safety Gate: The skill includes a 'Pre-Mutation Gate' that forbids modifying source files or permanent tests without affirmative human authorization (SKILL.md, Operating Invariants).
Audit Metadata