instrument-metric
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from the application's source code and outputs from Harness management tools to determine instrumentation logic, which introduces a surface for indirect instructions to influence the agent.
- Ingestion points: Local application source code files (Phase 2, Phase 4) and data returned by the
harness_gettool (Phase 5). - Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when interpolating retrieved code or tool data into the agent's context.
- Capability inventory: The agent has the capability to read and write files in the local repository (Phase 6) and execute Harness-specific MCP tools.
- Sanitization: No explicit validation or sanitization of the content extracted from the application codebase or the Harness API is mentioned before it is processed by the agent.
Audit Metadata