The Agent Skills Directory

[SAFE]: The skill exclusively provides instructional content for Git best practices, branch management, and Semantic Versioning (SemVer).
[SAFE]: It encourages security hygiene by providing patterns to check for secrets (passwords, API keys, tokens) in staged diffs before committing.
[SAFE]: It promotes the use of .gitignore to exclude sensitive files like .env and *.pem from version control.
[SAFE]: No suspicious external downloads, remote code execution, or data exfiltration patterns were identified. All shell commands provided (e.g., git worktree, git bisect) are standard developer tools used for their intended purposes.
[INDIRECT_PROMPT_INJECTION]: The skill describes workflows involving reading git logs, diffs, and blame information which could theoretically contain malicious instructions if a repository is compromised; however, this is an inherent property of version control systems and the skill provides no dangerous automated execution path for such content.
Ingestion points: Git history tools (git log, git diff, git blame).
Boundary markers: None specific; manual review is implied by the workflow.
Capability inventory: Git commands, standard Node.js development tools (npm, npx).
Sanitization: Not applicable as content is for human/agent review during development.

git-workflow-and-versioning