ai-image-editing

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface by processing untrusted user data—including images and natural language instructions—that is subsequently used to generate tool specifications and API calls.
  • Ingestion points: User-supplied source images and descriptive editing commands found in the interaction context.
  • Boundary markers: The TOUCH framework explicitly mandates human verification of results at 100% zoom and 'checking the seams' to ensure fidelity before hand-off or publishing.
  • Capability inventory: The skill is designed to invoke external image-editing APIs (e.g., FLUX, Firefly) and publish finalized content through the WoopSocial platform.
  • Sanitization: The instructions do not define programmatic sanitization, filtering, or the use of strict delimiters for external content within the agent's prompts.
  • [SAFE]: No findings related to credential exposure, malicious downloads, persistence mechanisms, or unauthorized privilege escalation were detected. The skill instructions prioritize legal compliance and user consent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 02:27 PM
Security Audit — agent-trust-hub — ai-image-editing