feedback-states
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires a connection to the vendor's MCP server at https://mcp.60fps.design/mcp to perform design searches and fetch motion code snippets using tools like 60fps_search_shots and 60fps_get_motion_code. These resources originate from the skill author's infrastructure.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes local SwiftUI and UIKit code to map application states, creating a potential attack surface for indirect prompt injection.
- Ingestion points: Processes local view models, network logic, and view files (SKILL.md, Step 1).
- Boundary markers: The instructions do not define specific delimiters to isolate local code from agent instructions.
- Capability inventory: The skill possesses the ability to call external MCP tools and generate executable code snippets (SKILL.md, Step 5).
- Sanitization: No explicit sanitization or filtering of local code content is mentioned.
Audit Metadata