meeting-synthesis
Pass
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed exclusively for text transformation, such as summarization and action item extraction, within the agent's context window.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted meeting data via the $ARGUMENTS variable.
- Ingestion points: Meeting notes and transcripts are ingested through the $ARGUMENTS field in SKILL.md.
- Boundary markers: The input is wrapped in tags to help the model distinguish between instructions and data.
- Capability inventory: The skill has no capabilities to execute code, access the file system, or make network requests.
- Sanitization: No explicit sanitization is performed, but the lack of dangerous tools minimizes the risk of exploitation.
Audit Metadata