dependency-flow
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection Defense: The skill proactively instructs the agent to treat data from external sources, such as GitHub issues or pull requests, as untrusted. This prevents the agent from executing commands found in comments or descriptions verbatim.
- Sensitive Command Guardrails: It establishes a clear list of prohibited operations (e.g., repository policy modification) and mandates explicit user confirmation for actions that trigger build promotion or modify dependency automation.
- Input Validation Procedures: The instructions define strict validation rules for version strings, build IDs, and channel names to ensure that data passed to internal tools is well-formed and legitimate.
- Controlled Tool Workflow: The skill structures the use of internal command-line and MCP tools by favoring read-only or lower-risk operations and providing MAUI-specific context to ensure they are used within their intended scope.
Audit Metadata