stocks-trend-screener-eval
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
git rev-parseto retrieve the current commit ID for logging purposes. This is a standard development workflow practice for tracking evaluation results against specific code versions. - [DATA_EXPOSURE]: The skill writes evaluation metrics and feedback to a local file (
.agents/skills/stocks-trend-screener/eval.csv). This is routine logging behavior for an evaluation tool and does not involve exfiltration of sensitive user data. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from
RUBRIC.mdandcases/*.mdto guide the evaluation subagents. While these files are external, they are part of the skill's own evaluation dataset. The skill manages boundaries by spawning isolated subagents for the 'actor' and 'judge' roles, which limits the scope of any potential injection in the test cases.
Audit Metadata