reel-builder
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions are standard and do not attempt to bypass safety filters. It exhibits an indirect prompt injection surface by ingesting untrusted data (teardowns and topics) into the prompt. Evidence chain: 1. Ingestion points:
SKILL.mdplaceholders; 2. Boundary markers: Absent; 3. Capability inventory: No command execution or network operations identified; 4. Sanitization: Absent. The risk is minimized as no exploitable capabilities are present in the skill code. - [CREDENTIALS_UNSAFE]: No hardcoded credentials detected. The skill provides explicit instructions to handle API keys securely through environment variables and warns against unauthorized file access for credential harvesting.
- [EXTERNAL_DOWNLOADS]: Mentions the use of well-known services like RunwayML and Remotion. No unverified or malicious remote code execution patterns or unsafe package installations were found.
- [SAFE]: The skill's behavior matches its stated purpose. No obfuscation, persistence mechanisms, or privilege escalation patterns were identified.
Audit Metadata