data-auditor-cleaner
Pass
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: SAFEPROMPT_INJECTIONREMOTE_CODE_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through the processing of untrusted data files. Evidence chain: 1. Ingestion points: Files are read from workspace/data_raw/ to map and audit data. 2. Boundary markers: The instructions lack delimiters or explicit warnings to ignore instructions embedded within the raw data content. 3. Capability inventory: The skill can read/write files and generate executable scripts. 4. Sanitization: No sanitization or escaping is specified for data interpolated into the agent's logic.
- [REMOTE_CODE_EXECUTION]: The skill performs dynamic script generation by creating cleaning code in workspace/code/scripts/ based on the structure and content of untrusted raw data. This allows external data inputs to potentially influence the logic of generated executable content.
Audit Metadata