md-report-summary
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill acts as a standard documentation and text-formatting assistant. It follows safe data handling practices, preserving original structure, images, and user tables without executing unsafe scripts or accessing sensitive files.
- [INDIRECT_PROMPT_INJECTION]: The skill has a potential surface for indirect prompt injection because it processes untrusted inputs.
- Ingestion points: External content is ingested via user-submitted drafts, text files, and
WebSearchresults (SKILL.md). - Boundary markers: Explicit XML/structural boundary markers are absent, though it includes text instructions to separate search results from original content.
- Capability inventory: Capabilities are limited to text processing and web searching; there are no dangerous execution capabilities like file modifications or code evaluation.
- Sanitization: The instructions mandate that the agent must not fabricate facts and should retain original data fields without modification.
Audit Metadata