commit
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHPROMPT_INJECTIONNO_CODE
Full Analysis
- Prompt Injection (HIGH): The skill is susceptible to indirect prompt injection attacks.
- Ingestion points: The skill ingests untrusted data from staged files.
- Boundary markers: There are no delimiters or explicit instructions to ignore embedded commands within the file content.
- Capability inventory: The skill metadata mentions 'potentially committing', which constitutes a write operation.
- Sanitization: No sanitization or validation of the external file content is performed prior to processing.
Recommendations
- AI detected serious security threats
Audit Metadata