The Agent Skills Directory

[SAFE]: The skill performs legitimate repository configuration tasks such as creating and moving markdown files for project management purposes. It follows best practices by confirming all changes with the user before execution.
[INDIRECT_PROMPT_INJECTION]: The skill reads existing repository files (e.g., CONTEXT.md, AGENTS.md) to detect current state. While this involves processing untrusted content from the repository, the skill's specific task list and the requirement for user review of the final output significantly mitigate the risk of an agent following instructions embedded in those files.

setup