try-fix
Pass
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- Command Execution: The skill utilizes PowerShell and Git to automate development workflows. It executes scripts located in the
.github/scripts/folder, such asBuildAndRunHostApp.ps1andEstablishBrokenBaseline.ps1. This allows the agent to perform builds and tests using the project's own established tooling.\n- File System Modification: To attempt fixes, the skill directly modifies target source files and creates a temporary logging structure underCustomAgentLogsTmp/. This is the intended operational behavior for a tool designed to empirically test code changes.\n- Indirect Prompt Injection Surface: The skill ingests external data like problem descriptions and test filters from SKILL.md. While this creates a potential surface for indirect injection, the instructions mandate the use of fixed repository scripts and a single-shot execution model, serving as a boundary for these inputs. The skill uses command execution and file-writing capabilities but relies on the repository's scripts to handle the sanitization of these parameters.
Audit Metadata