post-production-editing-agent
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill uses instructional headers like 'Core Rules (never break)' to define the agent's operational persona. These are standard formatting for agent instructions and do not exhibit malicious intent to bypass safety guardrails or override system instructions.
- [COMMAND_EXECUTION]: The skill references triggering 'ComfyUI + FFmpeg pipelines via Hermes RPC'. This is a functional description of how the agent interacts with authorized media processing tools and does not constitute unauthorized shell command execution.
- [SAFE]: The skill is purely instructional and does not contain hardcoded credentials, obfuscation, or persistence mechanisms. It possesses an ingestion surface for untrusted data via 'Campaign Briefs' and 'assets' (Ingestion points: SKILL.md; Boundary markers: Absent; Capability inventory: MP4 generation and FFmpeg RPC; Sanitization: Absent), but this is standard for media processing agents and no active risk was identified.
Audit Metadata