community-signal-digest

Pass

Audited by Gen Agent Trust Hub on May 5, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill uses search_external_content and commonroom_list_objects to fetch data from well-known community platforms. No unauthorized or suspicious network activity was detected.\n- [PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection by ingesting untrusted content from external platforms for use in automated drafting.\n
  • Ingestion points: Community posts and comments fetched via search tools and Common Room (SKILL.md Step 4, 5).\n
  • Boundary markers: Not specified for the data processing or drafting instructions.\n
  • Capability inventory: Capability to write to organizational intel records and generate drafts via dedicated writing skills.\n
  • Sanitization: Content is utilized for sentiment analysis and summary without explicit sanitization steps mentioned.\n- [EXTERNAL_DOWNLOADS]: Uses WebFetch to enrich signals with author profile data and view counts from well-known platforms like Stack Overflow and DEV.to (SKILL.md Step 8). This behavior is consistent with the primary purpose of community monitoring.
Audit Metadata
Risk Level
SAFE
Analyzed
May 5, 2026, 07:20 AM