The Agent Skills Directory

Automated Command Execution via Hooks: The skill utilizes a PostToolUse hook to trigger scripts/post-edit.sh automatically after file modifications. This script executes npx prettier to ensure consistent formatting. This is a common automation pattern for maintaining repository standards.
Local Script Integration: The workflow directs the agent to execute scripts/lint.sh for manual validation. Using repository-specific scripts is a standard practice for enforcing project-specific quality controls.
External Tooling and Dependencies: The formatting hook employs npx to run prettier. While npx can fetch packages from the public registry if not present locally, prettier is a well-recognized and trusted formatting utility.
Indirect Prompt Injection Surface: As the skill is designed to read and process documentation files, it naturally interacts with untrusted data which could contain embedded instructions.
Ingestion points: Markdown files (.md) read during the research and edit phases.
Boundary markers: No specific delimiters or "ignore" instructions are used when the agent reads file content.
Capability inventory: Git operations (branching, staging, committing), shell script execution (post-edit.sh, lint.sh), and external command execution via npx.
Sanitization: The skill does not perform content sanitization or validation of the text read from documentation files prior to editing.

write