The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes "Validate" commands specified in the project's tasks.md file. It also runs management scripts (fast-track-ai-gate.sh and fast-track-ai-index.sh) from sibling directories relative to the skill's installation. These scripts are recognized as author-provided resources.
[DYNAMIC_EXECUTION]: Shell commands are dynamically extracted from project documentation and executed at runtime, allowing for arbitrary execution based on file content.
[INDIRECT_PROMPT_INJECTION]: The skill uses untrusted documentation from the target repository to determine implementation steps and validation logic. * Ingestion points: Reads intent.md, spec.md, and tasks.md in the target repository. * Boundary markers: None. The skill treats the content of these files as trusted instructions for implementation and validation. * Capability inventory: Shell command execution, subagent task delegation, and file writing across the repository. * Sanitization: The skill does not perform sanitization on the commands or instructions extracted from the Markdown files.

ft-implement