story-film-eval
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The instructions describe running a local script (python scripts/run_evals.py) for static validation and suggest using a runner command for live models. These commands are intended for local development and testing of the skill package itself.
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to interpret and run evaluation cases, which involves processing data from external documentation and test files. This creates a potential surface for indirect prompt injection, although it is intended for testing such scenarios in a controlled developer environment.
- Ingestion points: Reads content from ../../references/EVALS.md and ../../evals/README.md.
- Boundary markers: None explicitly defined in the instruction text.
- Capability inventory: Execution of python scripts/run_evals.py and arbitrary runner commands.
- Sanitization: No evidence of sanitization or filtering of the processed evaluation data.
Audit Metadata