agency-clients

Pass

Audited by Gen Agent Trust Hub on Oct 24, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data from a PostEverywhere workspace, creating a potential surface for indirect instructions embedded in social media content.
  • Ingestion points: Data enters the context via list_accounts, list_campaigns, list_posts_advanced, and list_media calls as described in SKILL.md.
  • Boundary markers: The instructions include strong human-in-the-loop requirements, explicitly requiring the agent to 'ask the user for confirmation' before any destructive or creative action (Steps 3 and 4 in SKILL.md).
  • Capability inventory: The skill utilizes tools for account management, campaign creation, post scheduling (bulk_create_posts), and media retrieval (Steps 1, 5, and 7 in SKILL.md).
  • Sanitization: No specific technical sanitization is mentioned, though the skill relies on strict filtering logic (list_posts_advanced with specific IDs) to maintain data boundaries.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 24, 2026, 06:54 AM