The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute various local development commands such as building, linting, and type-checking (e.g., npm run build, npx tsc, ruff check). It also suggests the creation and execution of an init.sh script for environment health checks.
[EXTERNAL_DOWNLOADS]: The skill provides snippets for project configurations that include installing standard third-party development tools and packages from official registries (e.g., husky, lint-staged, ruff, pre-commit). It also references well-known public repositories for pre-commit hooks (e.g., astral-sh/ruff-pre-commit).
[PROMPT_INJECTION]: The skill implements a structured 'Initial Interview' and 'Sprint Contract' process designed to prevent 'Fuzzy Done' and other AI misalignment issues, which enhances the security and predictability of the agent's behavior.

harness-engineering