using-glean-productivity

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill logic is consistent with personal productivity synthesis using authorized vendor tools such as user_activity, meeting_lookup, and search.
  • [PROMPT_INJECTION]: The skill processes external data which constitutes an attack surface for indirect prompt injection.
  • Ingestion points: Untrusted data enters via email and document search results in reference/priorities.md and reference/activity.md.
  • Boundary markers: The output templates in reference/activity.md and reference/priorities.md lack delimiters to separate untrusted data from agent instructions.
  • Capability inventory: Analysis of SKILL.md shows capabilities are limited to retrieval tools; no command execution or file system write tools are present.
  • Sanitization: The workflow does not specify sanitization for the retrieved content before it is synthesized into summaries.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 06:51 PM
Security Audit — agent-trust-hub — using-glean-productivity