pr-review
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- Indirect Prompt Injection Surface: The skill is designed to ingest and analyze untrusted external content, such as PR descriptions, issue reports, and user comments. This information is used as context for the 'Try-Fix' phase, creating a potential path where malicious instructions in a PR or issue could influence the agent's code generation and testing behavior.
- Dynamic Code Implementation and Testing: In Phase 2, the agent implements fix ideas suggested by various AI models and executes test commands (e.g.,
dotnet test,Run-DeviceTests.ps1) to verify them. While this is the primary function of the skill, it involves running code that is generated at runtime based on external problem descriptions. - Autonomous Error Recovery: The skill provides instructions for the agent to handle environment blockers, such as missing tools or drivers, by attempting a single installation. This automated recovery, combined with the instruction to skip blocked phases without asking the user, highlights a high level of autonomy in managing the development environment.
Audit Metadata