rpi-walkthrough
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection Surface]: The skill processes external content like source code and UI definitions to provide explanations. As with any tool that reads project files, there is a potential for instructions embedded in those files to influence the agent's behavior. The skill mitigates this by focusing on descriptive walkthroughs rather than code execution.
- [State Management and File Access]: The skill creates and updates documentation within the
.copilot-tracking/directory to maintain a record of the walkthrough and any user-requested changes. This is an intended functional requirement for session persistence and does not involve unauthorized modification of the primary codebase. - [Subagent Orchestration]: The skill utilizes platform-provided subagents for exploration and research. This modular approach leverages existing agent capabilities for deep analysis while keeping the walkthrough logic focused on communication and documentation.
Audit Metadata