verification

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest and process data from external sources, including tool outputs and signed URLs, to verify project edits.
  • Ingestion points: SKILL.md (via preview_timeline, inspect_item, browse_assets, and inspect_asset).
  • Boundary markers: Absent; no specific delimiters or warnings for embedded instructions are provided to separate data from commands.
  • Capability inventory: SKILL.md (reading project structure, inspecting assets, and downloading URLs into a temporary directory).
  • Sanitization: Absent; no explicit validation or escaping of external content is defined in the instructions.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download media frames from signed URLs for visual verification. These resources originate from the vendor's plugin infrastructure as part of the intended verification workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 03:25 PM
Security Audit — agent-trust-hub — verification