The Agent Skills Directory

Privilege Escalation Guidance: The skill includes instructions that suggest using elevated permissions (sudo) to install system-level testing dependencies. While this is a common administrative task, it is a point of review for autonomous agents. Evidence in extensions/cpp.md: 'sudo apt-get install libgtest-dev catch2 gcovr lcov'.\n- External Resource Acquisition: The guidance describes processes for fetching external code and packages from public repositories and package managers. These references target well-known platforms like GitHub and official package registries, which are used to configure the test environment. Evidence in extensions/cpp.md: 'URL https://github.com/google/googletest/archive/refs/tags/v1.14.0.zip'.\n- Untrusted Data Processing Surface: The skill outlines a testing workflow that ingests data from local repositories using tools like git and find. This creates a surface where instructions embedded in the target codebase could potentially influence the agent's behavior. \n
Ingestion points: git status, git log, and directory scanning (in extensions/powershell.md).\n
Boundary markers: Not explicitly defined for repository-level metadata ingestion.\n
Capability inventory: The skill specifies capabilities for build execution, shell command execution, and package management across multiple languages.\n
Sanitization: No specific filtering or escaping mechanisms are described for the data ingested from version control logs or file paths.

code-testing-extensions