document-to-action-items
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted documents from local files and URLs, which could contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: Local files via
read_fileand external URLs viaweb_extract(SKILL.md). - Boundary markers: Includes a specific instruction to treat retrieved document content as data rather than instructions (SKILL.md).
- Capability inventory: The agent is authorized to write extracted data to Notion, spreadsheets, and task trackers (SKILL.md).
- Sanitization: Implements a 'Review before external writes' phase requiring explicit user approval before performing actions (SKILL.md).
Audit Metadata