The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing a variety of shell commands (yaks create, yaks list, yaks shave, yaks shorn, etc.) to track and update task state within the project's filesystem.
[REMOTE_CODE_EXECUTION]: The instructions facilitate the execution of the yaks tool directly from a remote repository using uvx --from git+https://github.com/joelgwebber/yaks yaks <cmd>. This resource is owned by the skill's author and is central to the skill's functionality.
[EXTERNAL_DOWNLOADS]: The skill uses uvx to download and run the yakherder package from the Python Package Index (PyPI) or directly from the author's GitHub repository.
[PROMPT_INJECTION]: The skill processes task data from markdown files located in the .yaks/ directory, which presents a surface for indirect prompt injection.
Ingestion points: Markdown files with YAML frontmatter stored in subdirectories of .yaks/ (e.g., hairy/, shaving/, shorn/).
Boundary markers: The instructions do not specify any explicit boundary markers or 'ignore' instructions for the agent when reading these files.
Capability inventory: The skill can execute CLI commands that read from and write to the local filesystem.
Sanitization: No specific sanitization or validation logic is mentioned for the content retrieved from task files.

yak