summarize
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content such as documents, reports, and conversation threads which could contain malicious instructions.\n
- Ingestion points: File paths, documents, and conversation history specified via
/summarizeor context.\n - Boundary markers: No specific delimiters or instructions to ignore embedded commands are defined.\n
- Capability inventory: The skill is restricted to text generation and does not define or utilize any dangerous capabilities like file writing, network access, or command execution.\n
- Sanitization: The skill contains an explicit security constraint in the 'Do not' section to not include secrets, credentials, or personal data found in the source material.
Audit Metadata