validation-scripts
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (HIGH): The skill is designed to ingest untrusted external data and process it using interpreters with execution capabilities. * Ingestion points: File './data/train.jsonl' read in 'examples/data-validation-example.md'. * Boundary markers: None provided to delimit data from instructions. * Capability inventory: Execution of 'bash' and 'python3'. * Sanitization: The main validation script is missing, leaving sanitization unverified.
- [Unverifiable Dependencies] (HIGH): The skill commands the execution of 'scripts/validate-data.sh', which is not included in the skill source, posing a risk of executing unverified code.
- [Dynamic Execution] (MEDIUM): Uses shell HEREDOCs to execute Python code at runtime in 'examples/data-validation-example.md'.
Recommendations
- AI detected serious security threats
Audit Metadata