The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Downloads and installs standard development dependencies from the official npm registry, including tools for linting, formatting, and git hook management.
[COMMAND_EXECUTION]: Executes shell commands via project package managers (npm, pnpm, yarn, or bun) to initialize tool configurations and perform migrations from legacy tools like ESLint or Prettier.
[COMMAND_EXECUTION]: Configures Git hooks in the .husky directory that execute local shell commands (e.g., gitleaks protect, lint-staged) during the commit workflow to enforce project standards.
[EXTERNAL_DOWNLOADS]: References the installation of the GitLeaks security utility via the well-known Homebrew package manager.
[PROMPT_INJECTION]: The skill ingests untrusted data from project files, creating a surface for indirect prompt injection.
Ingestion points: Reads package.json and project lockfiles (e.g., pnpm-lock.yaml, package-lock.json) to detect existing tooling and framework configurations.
Boundary markers: Absent; the skill scans file presence and names without explicit delimiters to isolate the data from core instructions.
Capability inventory: The skill utilizes shell execution via the Bash tool (scoped to package managers), as well as file modification via Write and Edit tools.
Sanitization: Absent; the skill relies on the structure of standard project metadata files to drive its decision-making logic.

setup