The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to read and process content from the project's codebase and documentation files, which is a standard ingestion surface for this type of task. 1. Ingestion points: The agent is instructed to read files within the packages/ and docs/ directories. 2. Boundary markers: While no explicit delimiters are defined to isolate external content, the agent's task is strictly scoped to documentation editing. 3. Capability inventory: The agent uses replace and write_file for file edits and suggests running npm run format for consistency. 4. Sanitization: Content is processed for writing tasks without specific technical sanitization, which is expected for a documentation agent.
[COMMAND_EXECUTION]: The skill includes a step to ensure project code style by suggesting a standard formatting command. 1. Evidence: The agent is instructed to offer to run npm run format once documentation changes are complete.

docs-writer