meeting-minutes
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted meeting transcripts and notes to generate structured meeting records and follow-up commands.
- Ingestion points: Meeting transcripts, notes, and conversation history are read as primary source material (SKILL.md, Step 2).
- Boundary markers: While the template (templates/meeting-minutes.md) organizes data into labeled sections, it does not implement explicit delimiters or safety instructions to prevent embedded commands within the untrusted text from being interpreted as valid project instructions.
- Capability inventory: The skill uses the Write tool to create artifacts, identifies the Bash tool as an allowed capability for project resolution, and generates shell-like commands for the user to execute (SKILL.md, Step 6).
- Sanitization: The instructions do not specify any validation, escaping, or filtering of the transcript content before it is included in the minutes or interpolated into follow-up commands.
Audit Metadata