autoresearch
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands to perform benchmarks and validation checks. This is the primary function of the optimization loop. \n- [DYNAMIC_EXECUTION]: The agent creates and executes local bash scripts (
autoresearch.sh) to run the optimization workload. \n- [INDIRECT_PROMPT_INJECTION]: The skill parses the output of benchmark scripts to extract performance metrics, which influences the agent's decision-making process. \n - Ingestion points: Standard output from
autoresearch.shandautoresearch.checks.sh. \n - Boundary markers: None present for benchmark output parsing. \n
- Capability inventory: Subprocess execution (bash), file system writes, and git operations. \n
- Sanitization: No explicit sanitization of benchmark output before parsing.
Audit Metadata