resemble-detect
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill communicates exclusively with the official Resemble AI API (app.resemble.ai) and adheres to standard security practices, such as using placeholders for API keys and HTTPS for all media transmissions.
- [SAFE]: No obfuscation, persistence mechanisms, or unauthorized privilege escalation patterns are present. The skill explicitly prevents the agent from guessing media authenticity, requiring a verified API result instead.
- [PROMPT_INJECTION]: An indirect prompt injection surface exists because the skill processes external media files and supports webhook callbacks, which could contain adversarial instructions. However, the risk is minimal given the skill's narrow scope and lack of high-risk capabilities.
- Ingestion points: Media file URLs, webhook callback URLs, and natural-language queries directed at completed detection results.
- Boundary markers: Not present; the skill does not use specific delimiters to separate untrusted data returned by the API from the system prompt.
- Capability inventory: Restricted to API requests to Resemble AI; no capabilities for local file system access, shell execution, or internal network operations.
- Sanitization: The skill does not describe specific sanitization for media intelligence results or transcriptions before they are presented to the agent.
Audit Metadata