agent-performance-analyzer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands in its pre and post lifecycle hooks to provide status feedback and record execution timestamps.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests performance metrics and execution traces, which represents a surface for indirect instructions if the source data is compromised.
- Ingestion points: Task execution monitoring and resource metric collection defined in SKILL.md.
- Boundary markers: Absent.
- Capability inventory: Shell execution in lifecycle hooks and use of the memory_store tool.
- Sanitization: Not specified for collected performance data.
Audit Metadata