summarize-meeting
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data in the form of meeting transcripts and screen activity retrieved via the
luciskill. - Ingestion points: Meeting transcripts and screen data captured from active windows.
- Boundary markers: The skill explicitly instructs the agent to read and follow user-defined rules in
~/.Life/RULES.md, which take precedence over other instructions. - Capability inventory: The skill has file-read and file-write capabilities restricted to the
~/.Lifedirectory structure. - Sanitization: No specific sanitization of the transcript content is mentioned; however, the output is constrained to a strict YAML frontmatter and Markdown format.
- [DATA_EXPOSURE]: The skill reads from
~/.Life/RULES.mdto determine processing logic and privacy constraints. This is a functional requirement for the skill's note-taking purpose and occurs within the skill's own data directory.
Audit Metadata