meeting-insights-analyzer

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external transcript files, which constitutes an indirect prompt injection surface.
  • Ingestion points: Reads meeting transcript files (.txt, .md, .vtt, .srt, .docx) from user-specified directories (e.g., ~/meetings/).
  • Boundary markers: The instructions do not define specific delimiters or explicit "ignore instructions" warnings for the agent to use when processing transcript content.
  • Capability inventory: The skill performs read-only analysis of text files and generates summaries; it lacks network access or file-write capabilities in its provided instructions.
  • Sanitization: There is no mention of sanitizing or escaping the content of the transcripts before analysis.
  • [NO_CODE]: The skill consists solely of markdown instructions and metadata for the AI agent; it does not include any external scripts, binaries, or executable code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 08:29 AM
Security Audit — agent-trust-hub — meeting-insights-analyzer