The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous configuration examples that execute shell commands (e.g., dotnet format, eslint, ruff, gitleaks) through git hook frameworks like Husky.Net and lefthook. These commands are localized to the user's development environment and are the intended primary function of the skill.
[EXTERNAL_DOWNLOADS]: The skill references several public GitHub repositories for use with the pre-commit framework, including well-known tools like astral-sh/ruff-pre-commit, pre-commit/pre-commit-hooks, and gitleaks/gitleaks. These are standard, reputable sources within the developer ecosystem.
[DATA_EXPOSURE_AND_EXFILTRATION]: No evidence of unauthorized data access or exfiltration. The skill includes examples for using GitHub Action secrets (${{ secrets.GITHUB_TOKEN }}) for CI/CD workflows, which is a standard and secure practice for credential management.
[INDIRECT_PROMPT_INJECTION]: As a tool for configuring git hooks, the skill creates an attack surface where untrusted data (such as commit messages or source code files from external contributors) is processed by automated tools.
Ingestion points: Frameworks (Husky, lefthook, pre-commit) read repository files and commit messages as documented in SKILL.md and examples/.
Boundary markers: The generated configurations do not include specific boundary markers or instructions to the AI agent to ignore embedded instructions in the processed data.
Capability inventory: The skill configures tools with subprocess execution capabilities (e.g., bash, npx, dotnet) as seen in examples/husky-net-dotnet-format.json.
Sanitization: Standard linter/formatter configurations are used without additional sanitization layers, which is consistent with the primary purpose of git hooks.

git-hooks