Enhance git-worktree-skills with official documentation and advanced patterns#9
Enhance git-worktree-skills with official documentation and advanced patterns#9tommy-ca wants to merge 8 commits intoEveryInc:mainfrom
Conversation
Introduce 3 reusable skills that encapsulate git worktree best practices: - git-worktree-create: Create worktrees with smart defaults - Branch naming conventions - Directory structure (.worktrees/) - .gitignore management - Verification steps - Includes automation script and examples - git-worktree-manage: List, remove, and maintain worktrees - Worktree lifecycle management - Cleanup and pruning - Status inspection - git-worktree-best-practices: Comprehensive workflow guide - Common patterns and workflows - Team conventions - Troubleshooting guidance These skills extract common patterns from commands (especially /work) to make them reusable across agents and other commands. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
Update the /work command to delegate worktree creation to the git-worktree-create skill instead of duplicating the logic inline. Changes: - Replace inline worktree setup (43 lines) with skill invocation - Reference skill documentation for details - Simplify command by ~14 lines - Maintain same functionality - Improve maintainability (single source of truth) The command now invokes the skill which handles: - Branch naming - Directory structure - .gitignore updates - Worktree creation - Environment verification 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
Release version 1.1.0 with new skills functionality. Changes: - Bump version: 1.0.0 → 1.1.0 (minor, additive change) - Update plugin.json: - Add components.skills: 3 - Add skills section with git-worktree skills - Update description to mention skills - Add git-worktrees keyword - Update marketplace.json: - Update version to 1.1.0 - Update description to mention skills - Add git-worktrees tag - Add plugin README.md: - Document all components (agents, commands, skills) - Include usage examples - List all agents, commands, and skills - Add version history - Add installation instructions This is an additive, non-breaking change that introduces reusable skills as a new capability of the plugin. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
Create a curated reference hub containing: - Official Git documentation (man pages, book, command references) - Community guides (DEV.to, Medium, GitHub Gists) - Tool integrations (CLI tools, IDE support) - Stack Overflow discussions - Related Git features (stash, sparse-checkout, bisect, rebase) This centralized resource makes it easier to find authoritative documentation and community best practices for git worktree workflows. Aligns with compounding engineering: create once, reference everywhere. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
…-create Add official Git documentation links to Overview with experimental status warning. Add "Git Worktree Fundamentals" section explaining: - Main vs linked worktrees - Shared vs isolated data (HEAD, index, files, refs) - Why this isolation matters for development Add "Advanced Worktree Options" section with: - Detached HEAD worktrees for experiments - Orphan branch worktrees for separate histories - Lock/unlock for portable worktrees - Repair operations for moved worktrees - Sparse checkout with no-checkout mode Add "Limitations and Warnings" section covering: - Experimental status with submodules - Branch checkout restrictions - Performance considerations with many worktrees - Best practices to avoid issues Add comprehensive "Official Resources" section with: - Official Git documentation references - Community guides and tutorials - Related skills and commands - Link to RESOURCES.md hub All changes reference official Git documentation. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
…manage Add official Git documentation links to Overview with command-specific references: - List, remove, prune, lock/unlock, repair commands Add "Worktree Status and Metadata" section explaining: - Status indicators (active, locked, prunable, detached HEAD) - Administrative files in .git/worktrees/ - What each metadata file represents Add "Advanced Management Techniques" section with: - Scripting with porcelain format output - Machine-readable output for parsing - Bulk operation examples - Repair operations for corrupted worktrees Add comprehensive "Official Resources" section with: - Official Git documentation references - Community guides and tutorials - Related skills - Link to RESOURCES.md hub All changes align with official Git documentation patterns. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
…ractices Add "Official Git Guidance" section with: - Key recommendations from Git documentation - Community best practices from GitHub Gist - Important experimental status warnings - Guidance on when and why to use worktrees Add Patterns 6-8 with real-world examples: - Pattern 6: Emergency Hotfix (canonical Git docs example) Shows worktrees as superior alternative to git stash - Pattern 7: CI/CD Optimization (single clone with worktrees) Saves disk space and bandwidth vs multiple clones - Pattern 8: Long-Running Review Workflow Preserves review context across days without branch switching Add comprehensive "Anti-Patterns to Avoid" section covering: - Too many worktrees (symptoms and solutions) - Manual directory deletion (use git worktree remove) - Shared branch in multiple worktrees (impossible) - Worktrees with submodules (experimental, use cautiously) Add "Compounding Engineering Philosophy" summary showing how worktrees enable: isolation → clean merges → easy cleanup → reduced cognitive load. All patterns reference official Git documentation and community best practices. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
skills/README.md: - Add "Learning Resources" section with links to: - Skill documentation files - Examples and usage patterns - RESOURCES.md curated reference hub - External official and community resources - Improves discoverability of comprehensive documentation plugins/compounding-engineering/README.md: - Add "Learning Resources" section with same structure - Helps users understand git-worktree-create, manage, and best-practices skills - Cross-references to skill documentation - Links to official Git docs and community guides - Mentions related `/work` and `/review` commands Makes 750+ lines of new documentation easily discoverable from top-level navigation points. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| log_info "Updating main branch" | ||
| cd "$git_root" | ||
| git checkout main 2>/dev/null || true | ||
| git pull origin main 2>/dev/null || true |
There was a problem hiding this comment.
Guard main-branch pull when checkout fails
The script always runs git pull origin main even if git checkout main failed. Because both commands are suffixed with || true, a checkout failure (for example due to uncommitted changes or a repository whose default branch is not main) is silently ignored while the pull still executes. In that scenario the pull merges remote main into whichever branch the user was already on, potentially introducing unrelated changes or conflicts when the intent was only to prepare a clean base before creating a worktree. Consider performing the pull only after a successful checkout or skipping the update when the checkout cannot be performed.
Useful? React with 👍 / 👎.
|
Closing this PR as it should be submitted to my fork (tommy-ca/every-marketplace) instead of the upstream repository. I will recreate this PR in my fork with the same changes. |
…-plan tracking Add the implementation plan for idea EveryInc#20 (Carrying Cost Budgeting) which introduces a `bun run skill:stats` command that ranks all skills and agents by carrying cost (`file_size x estimated_tool_calls`) instead of raw file size. The plan was refined through three parallel review passes (coherence, feasibility, scope) that applied 14 auto-fixes: - Collapsed two-file architecture to a single self-contained script - Broadened tool detection from verb+tool patterns (near-zero recall on real content) to capitalized tool names at word boundaries - Replaced loop scope-based multiplication with additive weights - Added `disable-model-invocation` frontmatter filter for tutorial skills - Clarified sort key as single `system_cost` metric (not two competing metrics) - Documented `@`-expansion undercount as a known limitation The meta-plan tracking table is updated: idea EveryInc#20 brainstorm and plan columns both marked done. Next in the roadmap: S4 (brainstorm ideas EveryInc#9, EveryInc#6, EveryInc#5, EveryInc#10 in parallel).
…-plan tracking Add the implementation plan for idea EveryInc#20 (Carrying Cost Budgeting) which introduces a `bun run skill:stats` command that ranks all skills and agents by carrying cost (`file_size x estimated_tool_calls`) instead of raw file size. The plan was refined through three parallel review passes (coherence, feasibility, scope) that applied 14 auto-fixes: - Collapsed two-file architecture to a single self-contained script - Broadened tool detection from verb+tool patterns (near-zero recall on real content) to capitalized tool names at word boundaries - Replaced loop scope-based multiplication with additive weights - Added `disable-model-invocation` frontmatter filter for tutorial skills - Clarified sort key as single `system_cost` metric (not two competing metrics) - Documented `@`-expansion undercount as a known limitation The meta-plan tracking table is updated: idea EveryInc#20 brainstorm and plan columns both marked done. Next in the roadmap: S4 (brainstorm ideas EveryInc#9, EveryInc#6, EveryInc#5, EveryInc#10 in parallel).
Add a conditional review persona that delegates to OpenAI's Codex CLI for an independent second-opinion read on the diff. Codex catches blind spots same-model reviewers share — different model, different reasoning patterns, different recall. Reapplies PR EveryInc#356 onto current main after the v3 restructure (ce-review -> ce-code-review, agents/<category>/ -> agents/ce-*.agent.md, ce: -> ce-, new subagent template that passes pre-computed diff). Addresses both Codex findings still active on the prior commit: - P1 (line 54): "Stop assuming ce:review passes base branch context." The new subagent template passes the pre-computed diff directly; ce-codex-reviewer materializes that diff to a tempfile and feeds it to codex. No base-branch resolution, no git symbolic-ref or rev-parse, no fall-back chain to break on non-default targets. - P2 (line 123): "Remove cross-persona suppression rule." The agent no longer instructs codex to suppress findings already covered by other personas. Synthesis dedupes centrally in Stage 5; the parallel subagent has no visibility into other reviewers' outputs anyway. Pieces: - agents/ce-codex-reviewer.agent.md — new persona, anchored confidence rubric (50/75; 100 never emitted because second-opinion is not direct verification), two-phase environment guard (CODEX_SANDBOX/SESSION_ID recursion check + which codex availability check), pipe-delimited output translation into the structured findings schema. - skills/ce-code-review/references/persona-catalog.md — codex registered as conditional persona EveryInc#9. Selected when the diff is non-trivial (>=25 changed non-test lines) or correctness-sensitive (auth, payments, data, parsing, concurrency, external API contracts) AND the user-level codex CLI is on PATH. - skills/ce-code-review/SKILL.md — codex added to the cross-cutting conditional dispatch table; reviewer count bumped 18 -> 19. - plugins/compound-engineering/README.md — codex listed in the Review agents table (insertion in sorted order). - tests/review-skill-contract.test.ts — codex added to the anchored- rubric persona loop, plus four new contract tests guarding (a) frontmatter shape, (b) the no-base-branch-resolution rule with explicit P1 regression callouts, (c) the no-cross-persona-suppression rule with explicit P2 regression callout, (d) catalog/SKILL.md registration, and (e) the two environment guards before any codex invocation. Verification: - bun test tests/review-skill-contract.test.ts: 28 pass (was 24) - bun test full suite: 992 pass, 9 pre-existing failures in detect-project-type.sh unchanged (present on main before this PR) - bun run release:validate: clean (52 agents, 35 skills, 0 MCP) No release-owned files touched: plugin.json, marketplace.json, and CHANGELOG.md remain untouched per AGENTS.md rules.
Summary
Enhanced the three git-worktree-skills with comprehensive references to official Git documentation, community best practices, and advanced workflow patterns. Created a centralized resource hub and organized changes into 5 atomic, logically independent commits for clean code review.
What's Changed
New Files
Enhanced Skills
git-worktree-create
git-worktree-manage
git-worktree-best-practices
Documentation Updates
Atomic Commits
This PR contains 5 logically independent commits:
Add centralized RESOURCES.md (e0dc644)
Add documentation, fundamentals, and advanced options to git-worktree-create (d284523)
Add documentation, metadata, and advanced management to git-worktree-manage (c8e1d95)
Add official guidance, workflow patterns, and anti-patterns to best-practices (568a1de)
Add Learning Resources sections to README files (b27bd0d)
Testing
✓ Verified all code examples execute correctly in test repository
✓ Tested worktree creation, listing, and removal operations
✓ Validated markdown formatting and structure
✓ Confirmed all JSON files valid
Content Stats
Compounding Engineering Impact
Following the compounding engineering philosophy:
Related Issues
This enhancement addresses the need for:
Review Notes
Each commit is:
The structure allows reviewing by logical concern rather than by file.
🤖 Generated with Claude Code