The Agent Skills Directory

[SAFE]: The skill instructions emphasize security best practices for repository management. It specifically instructs the agent to never bypass git hooks ("--no-verify") and to never force-push to branches, maintaining the integrity of the project's safety net.
[COMMAND_EXECUTION]: The skill uses local CLI tools including "git" and "gh" for standard branching operations, merging, and release creation. These commands are executed within the context of the user's repository and are required for the workflow described.
[PROMPT_INJECTION]: The skill involves a surface for indirect prompt injection as it processes data from git history to generate changelogs. Ingestion points: External data enters through "git log" and PR subjects processed in "references/changelog.md". Boundary markers: Absent; there are no specific markers used to delimit data from instructions when reading history. Capability inventory: The agent can perform "git push", "git merge", and "gh release create" (detailed in "references/release.md" and "references/hotfix.md"). Sanitization: Absent; however, the data is primarily used for textual documentation, which limits the potential for exploitation.

gitflow