skill-benchmark
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes standard shell utilities (
find,wc,cat) to analyze project structure and read configuration files likeAGENTS.md. These commands are used for introspection and do not perform any destructive actions or network exfiltration. - [DATA_INDIRECT_INJECTION]: The skill ingests external project data, including source code and evaluation JSON files, to perform its refactoring tasks. This creates a potential surface for indirect prompt injection if the ingested files contain malicious instructions.
- Ingestion points: Project source files (
src/*.ts,src/*.tsx),AGENTS.md, andevals/evals.json. - Boundary markers: No specific boundary markers or 'ignore' instructions are used when processing the contents of these files.
- Capability inventory: The agent performs file system reads and provides refactoring suggestions based on the ingested data.
- Sanitization: No sanitization or validation of the external file content is performed before it is used to influence the agent's logic.
Audit Metadata