The Agent Skills Directory

[SAFE]: The skill establishes safe operational guidelines for Git, specifically prohibiting the bypass of commit hooks and requiring user approval for all commits, merges, and rebases.
[NO_CODE]: The skill is purely instructional and contains no scripts or binaries, minimizing the direct execution surface.
[PROMPT_INJECTION]: The workflow involves reading and summarizing Git history for releases (Section 5), which is an indirect prompt injection surface. An attacker could attempt to influence the agent via commit messages, but this is mitigated by the required manual review of the generated summaries.
Ingestion points: Git commit history (SKILL.md, Section 5)
Boundary markers: Absent
Capability inventory: Shell access via the git command-line tool
Sanitization: Absent

git-workflow