revyl-mcp
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to interact with live mobile applications and process UI state via screenshots and MCP tools, which inherently creates an attack surface for indirect prompt injection from application content.
- Ingestion points: Application UI state and visual data processed via
screenshot()and various MCP tool calls referenced in the orchestration rules. - Boundary markers: None identified; the skill does not specify delimiters or instructions to ignore embedded content within the application being tested.
- Capability inventory: Mentions
screenshot()and coordinates calls to other specialized skills likerevyl-mcp-dev-loopfor live interaction. - Sanitization: No explicit sanitization or filtering of application content is defined within these instructions.
- [SAFE]: The skill functions as a configuration and routing layer for Revyl's mobile testing suite. All referenced resources and skills (e.g.,
revyl-mcp-dev-loop,revyl-cli) are consistent with the vendor's intended functionality and do not exhibit suspicious behavior.
Audit Metadata