The Agent Skills Directory

[PROMPT_INJECTION]: The skill identifies the project stack by reading local files, creating a surface for indirect prompt injection. • Ingestion points: Reads marker files like package.json, pyproject.toml, and tsconfig.json from the project workspace. • Boundary markers: No explicit markers are used to delimit untrusted project data. • Capability inventory: The skill can write configuration files and execute shell commands via package managers. • Sanitization: Project file content is used to drive logic without explicit validation.
[COMMAND_EXECUTION]: The skill generates and suggests shell commands (e.g., npm install, ruff check) for both local environments and automated CI/CD workflows.
[EXTERNAL_DOWNLOADS]: The skill downloads standard utilities and GitHub Actions from trusted organizations and official registries (NPM, PyPI, and GitHub).

code-quality