From coding
Plan and execute coherent Conventional Commit groupings for large, tangled working tree changes across many files.
How this skill is triggered — by the user, by Claude, or both
Slash command
/coding:gitplanThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
**Don't invoke this skill if we're just reviewing a couple files. This skill is ONLY applicable when untangling 15+ files with multiple inter-twingled edits.**
Don't invoke this skill if we're just reviewing a couple files. This skill is ONLY applicable when untangling 15+ files with multiple inter-twingled edits.
Never create megacommits. Each commit should be focused, coherent, and reviewable.
If the repository has a layered structure (e.g. shared utilities → core → feature packages → app), work through it from the lowest-level layer upward so dependencies land before their consumers.
git status and git diff --stat — use subagents to preserve context and give summaries.git diff -U150 but limit JSON/lockfiles to the first ~50 lines.git add <files>, including any related docs/plans decided in Phase 1.review-staged skill if available). Use a capable model — reviews are important.git status for remaining changes.Review the staged code for potential issues and improvements. Follow any project conventions documented in the repository root (e.g. CLAUDE.md, AGENTS.md, contributing guides).
For each issue:
For documentation or trivial implementation issues, suggest the edit to the user and apply if they accept.
For other issues, provide a unique identifier for each issue (e.g. #A or #B), a summary of the issue, where it's located, and a proposed solution. Ask the user and apply if they accept.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub photostructure/claude-code-skills --plugin coding