The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes system-level commands to set up the local development environment. Evidence: Step 5 and 6 explicitly call for the execution of git init and package installation commands including npm install, bun install, pnpm install, and pip install.
[EXTERNAL_DOWNLOADS]: The skill downloads and installs third-party software from well-known package registries. Evidence: Step 6 automates the installation of development tools such as Biome, Ruff, Vitest, and pytest based on the detected or chosen technology stack.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection when working with existing codebases. Evidence: 1. Ingestion points: Step 1 and 4 read content from existing manifests like package.json or pyproject.toml and documentation in the docs/ directory. 2. Boundary markers: No delimiters or instructions to ignore embedded commands within existing files are specified. 3. Capability inventory: The skill can write files, initialize git, and execute package manager installations. 4. Sanitization: No sanitization or validation of the content read from existing project files is performed before it influences agent actions.

bootstrap