git-commits
The git-commits skill instructs Claude Code to use a dedicated /commit function instead of running git commit directly, which removes Claude attribution metadata from commits and generates a reasoning.md file documenting the work attempted. Use this skill whenever a user requests to commit, push, or save changes to ensure commits appear user-authored and maintain a development history for future sessions.
git clone --depth 1 https://github.com/parcadei/Continuous-Claude-v3 /tmp/git-commits && cp -r /tmp/git-commits/.claude/skills/git-commits ~/.claude/skills/git-commitsSKILL.md
# Git Commit Rules
When the user asks to commit, push, or save changes to git:
## MUST Use /commit Skill
**DO NOT** run `git commit` directly. Instead:
```
Skill("commit")
```
The `/commit` skill:
1. Removes Claude attribution from commits
2. Generates reasoning.md capturing what was tried
3. Clears build attempts for next feature
## Why This Matters
- Regular `git commit` adds "Generated with Claude Code" and Co-Author lines
- The `/commit` skill removes these so commits appear user-authored
- Reasoning capture preserves build history for future sessions
## Trigger Words
When you see these in user prompts, use the commit skill:
- "commit", "push", "save changes"
- "push to github", "push changes"
- "commit and push"
## After Commit
The skill will prompt you to run:
```bash
bash "$CLAUDE_PROJECT_DIR/.claude/scripts/generate-reasoning.sh" <hash> "<message>"
```
Then push if requested:
```bash
git push origin <branch>
```Security vulnerability analysis and testing
Build Python agents using Agentica SDK - spawn agents, implement agentic functions, multi-agent orchestration
Unit and integration test execution and validation
Feature planning, design documentation, AND integration planning
End-to-end and acceptance test execution
Analyze Claude Code sessions using Braintrust logs
Session analysis, precedent lookup, and learning extraction
Query the artifact index for precedent and guidance