dataset-curator
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or unauthorized behaviors were detected in the skill configuration. The tool performs legitimate data curation and privacy auditing as described in the documentation.
- [PROMPT_INJECTION]: The skill processes untrusted data files, which is a known surface for indirect prompt injection. Ingestion points: The input argument accepts JSON, CSV, and text files for curation. Boundary markers: Not explicitly specified in the configuration. Capability inventory: The skill reads from the local filesystem and writes output files. Sanitization: The skill includes built-in PII auditing and noise removal to sanitize datasets before use.
Audit Metadata