The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to perform various repository management and file system operations to maintain the skills index.
Evidence:
Execution of Git commands including git clone, git status, git diff, and git push to synchronize changes with a remote repository (e.g., L-yifan/skills).
File system operations such as directory copying and path resolution to manage skill sources.
Automated deletion of temporary clone directories and intermediate source copies following successful synchronization, with built-in path verification to ensure safety.
[INDIRECT_PROMPT_INJECTION]: The skill ingests content from external, third-party skills to update its README index, creating a surface for potential injection.
Ingestion points: The skill reads SKILL.md files and official installation instructions from external GitHub repositories or marketplace links (Step 2A).
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined for the imported external text.
Capability inventory: The skill possesses file-write capabilities (updating the README and syncing directories) and network access (git push).
Sanitization: The instructions focus on formatting consistency within the README tables but do not explicitly describe filtering or sanitization of the external content.

skill-backup