The Agent Skills Directory

Command Execution: The skill is designed to automate repository-specific workflows by executing make targets and git commands. This allows it to run formatting, linting, and test suites as defined in the project's own configuration files.
Dynamic Script Generation: The PowerShell implementation (run.ps1) includes a mechanism to generate a temporary wrapper script at runtime. This logic is used to ensure the make command executes correctly in environments where it might be defined as a shell function rather than an external application.
Indirect Prompt Injection Surface: The skill evaluates external environment state, specifically host process and task evidence, to manage execution timing and avoid contention (ingestion point identified in SKILL.md). The capability inventory for this logic includes shell commands such as make, git, and process cleanup tools in run.sh and run.ps1. While no specific boundary markers or sanitization routines are defined for this telemetry, the risk is mitigated by its specialized use for scheduling and the absence of complex data processing.
Execution Policy Bypass: The instructions for Windows users include the -ExecutionPolicy Bypass flag for PowerShell. This is a common administrative setting used to facilitate the execution of local scripts in development environments.
External Resource Access: The skill references the official Python Package Index (PyPI) for dependency management via the UV_DEFAULT_INDEX environment variable. This targets a well-known service for standard package resolution.
Credential Protection: As a security measure, the run.sh script explicitly unsets the OPENAI_API_KEY environment variable before starting the verification processes. This practice helps ensure that sensitive credentials are not accessible to the test environment or child processes.

code-change-verification