meeting-notes
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were detected. The skill uses standard markdown templates and logical instructions to process user-provided meeting data. It does not perform any network operations, access sensitive files, or execute external code. The indirect prompt injection surface is documented below for completeness, but it is considered low risk given the skill's purpose.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted meeting transcripts or notes provided by the user.
- Ingestion points: Raw text, transcripts, or chat logs provided in the user prompt (SKILL.md, Step 1).
- Boundary markers: The instructions guide the agent to structure the output using specific markdown headers and tables, which helps keep the agent's output focused.
- Capability inventory: The skill only performs text transformation and formatting. There are no subprocess calls, file-write operations, or network requests.
- Sanitization: None present; the skill assumes the input is a meeting transcript. However, the transformation into structured templates (tables and specific headers) inherently limits the potential for an embedded instruction in the transcript to divert the agent's behavior.
Audit Metadata