The Agent Skills Directory

[SAFE]: No security issues detected. The skill is designed with strong internal controls and follows least-privilege principles by restricting platform interactions to the official uip CLI and forbidding direct API or credential access.
[COMMAND_EXECUTION]: The skill executes uip CLI commands to interact with the UiPath platform and standard Windows system commands like sc query and tasklist for local diagnostics. These operations are aligned with the skill's primary purpose of diagnosing automation and platform issues.
[EXTERNAL_DOWNLOADS]: The skill retrieves technical documentation and troubleshooting guides from UiPath's official knowledge base using the uip docsai ask tool. These references target trusted vendor resources.
[DATA_EXPOSURE]: The skill accesses local robot logs and project source code to identify the root cause of workflow failures. This data is processed locally within the .investigation/ directory and summaries are interpreted by specific sub-agents.
[PROMPT_INJECTION]: Heuristic detections of instruction concealment in the 'Healing Agent Data' reference are false positives. The instructions provided are intended to ensure user-friendly presentation by translating technical JSON fields into natural language, rather than hiding malicious actions from the user.

uipath-diagnostics