The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches content (skills and documentation) from the author's GitHub repository (ccheever/llp). This is a standard part of the skill's function to manage installations and updates.
[COMMAND_EXECUTION]: The workflow requires the use of shell commands, specifically git for operations such as shallow cloning, archiving, and reverting changes during the update process.
[INDIRECT_PROMPT_INJECTION]: The skill processes the contents of the local repository to generate documentation drafts. It implements strong mitigation strategies by:
Ingestion points: Surveying the codebase history and structure.
Boundary markers: Using provenance tags like [observed], [confirmed], and [inferred] to distinguish facts from AI assumptions.
Capability inventory: Writing files to the llp/ directory, managing AGENTS.md, and executing git commands.
Sanitization: Requiring mandatory human diff-approval before any files are overwritten or promoted from Draft status.

llp-adopt