The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes structured configuration data from the user's repository, creating a surface for indirect prompt injection.
Ingestion points: The agent reads turbo.json, package.json, and .env files to identify task pipelines and environment configurations (see SKILL.md and command/turborepo.md).
Boundary markers: No specific boundary markers or instructions to disregard natural language commands embedded in these files are provided.
Capability inventory: The skill enables the agent to execute turbo commands and other monorepo tools (e.g., npx turbo-ignore) using the command/turborepo.md workflow.
Sanitization: No content validation is performed to filter out potentially malicious instructions in the processed repository files.
[EXTERNAL_DOWNLOADS]: Fetches and executes the turbo-ignore tool from the npm registry using npx during Vercel deployment workflows. This reference to Vercel's official utility is documented neutrally.
[COMMAND_EXECUTION]: Orchestrates various monorepo operations using the Turborepo CLI, including turbo run, turbo watch, and turbo boundaries.

turborepo