error-debugging-error-trace
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- NO_CODE (SAFE): The analyzed skill consists solely of a markdown instruction file (SKILL.md) and does not include any Python scripts, JavaScript, or shell executables.
- PROMPT_INJECTION (SAFE): No malicious override or bypass patterns were detected. The instructions follow standard system prompting for a specialized persona.
- DATA_EXFILTRATION (SAFE): The skill does not perform network operations or access sensitive system paths. It explicitly instructs the agent to 'Avoid logging secrets, tokens, or personal data'.
- INDIRECT_PROMPT_INJECTION (LOW): The skill is designed to process error traces and logs, which are untrusted data sources.
- Ingestion points: Error capture and grouping workflows mentioned in the 'Instructions' section.
- Boundary markers: Not explicitly defined in the prompt text.
- Capability inventory: The skill is limited to configuration and assessment advice; no automated tool execution or file-writing capabilities are provided in the source.
- Sanitization: The 'Safety' section provides basic guidance to prevent secret logging.
Audit Metadata