The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies extensively on shell commands to manage the repository, including git operations for branching, merging, and logging.
[COMMAND_EXECUTION]: It executes npm run build and npm test to verify the project's integrity after merging upstream changes.
[COMMAND_EXECUTION]: Provides instructions to the user for restarting services using launchctl on macOS platforms.
[EXTERNAL_DOWNLOADS]: Connects to remote git repositories to fetch updates, with a default configuration targeting the author's repository on GitHub.
[PROMPT_INJECTION]: The skill parses CHANGELOG.md to identify breaking changes and offers to execute associated migration skills. This represents an indirect prompt injection surface as the agent's behavior is influenced by content in the repository's documentation.
Ingestion points: Step 6 reads and parses CHANGELOG.md content following a merge.
Boundary markers: None; the logic looks for specific text patterns ([BREAKING]) within the file.
Capability inventory: The skill can invoke other skills and execute arbitrary shell commands via git and npm.
Sanitization: No validation is performed on the skill names extracted from the changelog before offering them for execution.

update-nanoclaw