The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes project-local commands pnpm docs:check and pnpm docs:build to validate documentation integrity and verify builds. These commands are standard for the documentation framework used (Astro/Starlight).
[INDIRECT_PROMPT_INJECTION]: The skill involves reading and editing markdown and MDX files from the repository, which presents a theoretical surface for indirect prompt injection if those files contain malicious instructions meant to influence the agent.
Ingestion points: Documentation content is read from packages/docs/src/content/docs/** and configuration from packages/docs/astro.config.mjs.
Boundary markers: No specific delimiters or instructions to ignore embedded commands in the documentation data are provided.
Capability inventory: The agent has the capability to read and write files and execute local pnpm shell commands.
Sanitization: No explicit sanitization or filtering of the ingested documentation content is described.

public-docs