perf-investigation-logger
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and store verbatim user quotes into performance investigation logs.
- Ingestion points: User-provided quotes are processed and written to markdown files in the
{state-dir}/perf/investigations/directory. - Boundary markers: The output format uses standard double quotes for the user quote section.
- Capability inventory: The skill performs filesystem write operations (appending to markdown files).
- Sanitization: There is no explicit sanitization of the user-provided text, which is expected for a verbatim logging utility. While this creates a surface where malicious instructions could be stored in logs, the risk is minimal given the specialized context of performance debugging.
Audit Metadata