analyze-document-library
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill follows best practices by explicitly enforcing a read-only policy and forbidding any state-changing operations.\n- [PROMPT_INJECTION]: The skill processes file and folder names from external sources, which represents a potential surface for indirect prompt injection.\n
- Ingestion points: Library contents are retrieved in Step 2 from the active SharePoint context.\n
- Boundary markers: The instructions include explicit negative constraints (e.g., 'Do not modify', 'Do not rename') and a strict output format to maintain the agent's focus on analysis.\n
- Capability inventory: No tools for file system modification, network exfiltration, or code execution are utilized in the skill.\n
- Sanitization: The skill relies on structured output and negative instructions rather than explicit string sanitization.
Audit Metadata