error-detective
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is purely instructional and does not contain any executable scripts, binaries, or commands that could be used for malicious purposes.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze logs and stack traces, which are untrusted external inputs. This creates a surface for indirect prompt injection where instructions hidden in logs could attempt to influence the agent. However, the skill explicitly provides instructions to mitigate this risk.
- Ingestion points: Logs, stack traces, and codebase snippets provided by the user (SKILL.md).
- Boundary markers: No specific delimiters (like triple backticks or XML tags) are mandated for the log data in the instructions.
- Capability inventory: The skill focuses on analysis, regex generation, and manual query construction; it does not authorize the agent to perform instrumentation or deployment.
- Sanitization: The skill mandates that the agent must redact tokens, personal data, and session content before sharing logs, which follows security best practices for log handling.
Audit Metadata