The Agent Skills Directory

[COMMAND_EXECUTION]: The skill implements a tiered safety model for executing Git and GitHub commands via the gw wrapper, including mandatory flags for write operations and strict protection for critical branches.
[DATA_EXFILTRATION]: GitHub API interactions and remote operations are limited to standard repository management. References to the 'AutumnsGrove/Lattice' repository are associated with the verified vendor.
[PROMPT_INJECTION]: The skill provides clear usage guidelines for Git workflows and defines data ingestion surfaces for batch issue creation as standard administrative tasks, without attempting to bypass safety guardrails.

git-workflows