meeting-briefing

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • Indirect Prompt Injection Surface: The skill is designed to ingest and process data from various external sources, which is a common pattern for assistants but introduces a potential surface for indirect instructions.
  • Ingestion points: In SKILL.md, the agent is instructed to pull information from Calendar, Email, Chat (Slack/Teams), Document storage (Box/Egnyte/SharePoint), CLM, and CRM systems.
  • Boundary markers: The instructions do not include specific delimiters or guidance to the agent to disregard instructions that might be embedded within those external data sources.
  • Capability inventory: The skill is primarily instructional and does not include scripts that execute subprocesses, write files, or perform network operations outside of the platform's standard data fetching.
  • Sanitization: There is no evidence of specific sanitization or filtering logic applied to the external data before it is interpolated into the meeting briefing synthesis.
  • No Code Components: The skill consists entirely of markdown instructions without accompanying scripts, binary executables, or package manifests, which simplifies the security profile.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:16 AM
Security Audit — agent-trust-hub — meeting-briefing