@uniswap/ai-toolkit-nx-claude 0.5.29 → 0.5.30-next.0

package/dist/content/commands/agnostic/claude-init-plus.md

@@ -1,439 +0,0 @@
----
-description: Initialize or update CLAUDE.md files at all core nodes in the workspace
-argument-hint: (no arguments - scans entire workspace automatically)
-allowed-tools: Glob(*), Read(*), Write(*), Edit(*), Bash(tree:*), Bash(find:*), Bash(git:*)
----
-
-# `/claude-init-plus` - Intelligent CLAUDE.md Generation
-
-## Purpose
-
-Scan the workspace to identify "core nodes" and generate/update CLAUDE.md files at each node. Works with any workspace type: monorepos (Nx, Turborepo, Lerna, npm/yarn/pnpm workspaces), single-repo projects, or any language ecosystem.
-
-This creates documentation that helps future Claude Code sessions understand the project without reading every file.
-
-## Core Node Definition
-
-A directory is a **core node** if it meets **either** criterion:
-
-1. **Has a package manifest** → Always a core node
-
-   - JavaScript/TypeScript: `package.json`
-   - Rust: `Cargo.toml`
-   - Python: `pyproject.toml`, `setup.py`
-   - Go: `go.mod`
-   - Java: `pom.xml`, `build.gradle`, `build.gradle.kts`
-   - Swift: `Package.swift`
-   - Ruby: `Gemfile`, `.gemspec`
-   - PHP: `composer.json`
-
-2. **Is a working directory**:
-   - Contains ONLY files (no subdirectories), OR
-   - Contains files + subdirectories where ALL subdirectories contain ONLY files (max 1 nesting level)
-
-**Examples:**
-
-- ✅ `apps/web/` (has package.json)
-- ✅ `backend/api/` (has Cargo.toml)
-- ✅ `.github/workflows/` (only files)
-- ✅ `src/components/` (files + shallow subdirs with only files)
-- ❌ `apps/` (container with multi-level dirs)
-- ❌ `.git/` (deep structure)
-- ❌ `node_modules/` (excluded by default)
-
-## Usage
-
-```bash
-/claude-init-plus
-```
-
-## Execution Flow
-
-### Phase 1: Discovery
-
-**Step 1: Use Git for Fast Discovery (Preferred)**
-
-```bash
-# Get all files tracked by git (automatically excludes ignored files)
-git ls-files
-
-# Find all package manifests
-git ls-files | grep -E '(package\.json|Cargo\.toml|pyproject\.toml|setup\.py|go\.mod|pom\.xml|build\.gradle|Package\.swift|Gemfile|.*\.gemspec|composer\.json)$'
-```
-
-**Fallback: Use find if not a git repo**
-
-```bash
-find . -type f \
-  -not -path '*/node_modules/*' \
-  -not -path '*/.git/*' \
-  -not -path '*/dist/*' \
-  -not -path '*/build/*' \
-  -not -path '*/.nx/*' \
-  -not -path '*/.cache/*' \
-  -not -path '*/target/*' \
-  -not -path '*/__pycache__/*' \
-  -not -path '*/venv/*' \
-  -not -path '*/.venv/*'
-```
-
-**Step 2: Identify Core Nodes Using BFS**
-
-Traverse workspace using BFS (breadth-first search):
-
-```
-For each directory:
-  1. Check exclusion list (skip if matches)
-  2. Apply core node detection:
-
-     RULE 1 (HIGHEST PRIORITY): Check for package manifest
-     if [ -f "$dir/package.json" ]; then
-       → Mark as Core Node (Package: npm-package, JavaScript/TypeScript)
-       → Read manifest for name, scripts, dependencies
-       → Skip to next directory
-     elif [ -f "$dir/Cargo.toml" ]; then
-       → Mark as Core Node (Package: cargo-package, Rust)
-       → Read manifest for name, dependencies
-       → Skip to next directory
-     elif [ -f "$dir/pyproject.toml" ]; then
-       → Mark as Core Node (Package: python-package, Python)
-       → Read manifest for name, scripts, dependencies
-       → Skip to next directory
-     elif [ -f "$dir/setup.py" ]; then
-       → Mark as Core Node (Package: python-package, Python)
-       → Read setup.py for basic info
-       → Skip to next directory
-     elif [ -f "$dir/go.mod" ]; then
-       → Mark as Core Node (Package: go-module, Go)
-       → Read go.mod for module name
-       → Skip to next directory
-     elif [ -f "$dir/pom.xml" ]; then
-       → Mark as Core Node (Package: maven-project, Java)
-       → Read pom.xml for artifactId, groupId
-       → Skip to next directory
-     elif [ -f "$dir/build.gradle" ] || [ -f "$dir/build.gradle.kts" ]; then
-       → Mark as Core Node (Package: gradle-project, Java/Kotlin)
-       → Read build.gradle for project info
-       → Skip to next directory
-     elif [ -f "$dir/Package.swift" ]; then
-       → Mark as Core Node (Package: swift-package, Swift)
-       → Read Package.swift for package name
-       → Skip to next directory
-     elif [ -f "$dir/Gemfile" ] || [ -f "$dir"/*.gemspec ]; then
-       → Mark as Core Node (Package: ruby-gem, Ruby)
-       → Read Gemfile or .gemspec for gem info
-       → Skip to next directory
-     elif [ -f "$dir/composer.json" ]; then
-       → Mark as Core Node (Package: composer-package, PHP)
-       → Read composer.json for name, scripts
-       → Skip to next directory
-     fi
-
-     RULE 2: Check if working directory
-     files=( "$dir"/* )           # Get all files
-     subdirs=( "$dir"/*/ )        # Get all subdirectories
-
-     # Case A: Only files, no subdirectories
-     if [ ${#subdirs[@]} -eq 0 ] && [ ${#files[@]} -gt 0 ]; then
-       → Mark as Core Node (Working Directory - Leaf)
-       → Record type: "working-dir-leaf"
-     fi
-
-     # Case B: Files + shallow subdirectories (1 level max)
-     if [ ${#files[@]} -gt 0 ] && [ ${#subdirs[@]} -gt 0 ]; then
-       all_subdirs_are_leaves=true
-       for subdir in "${subdirs[@]}"; do
-         if find "$subdir" -mindepth 1 -type d | grep -q .; then
-           all_subdirs_are_leaves=false
-           break
-         fi
-       done
-
-       if [ "$all_subdirs_are_leaves" = true ]; then
-         → Mark as Core Node (Working Directory - 1-Level)
-         → Record type: "working-dir-nested"
-       fi
-     fi
-
-  3. If not a core node, continue traversing into subdirectories
-```
-
-**Exclusion List:**
-`.git`, `node_modules`, `.nx`, `.expo`, `dist`, `build`, `.next`, `.vercel`, `.yarn`, `coverage`, `.turbo`, `.cache`, `logs`, `target`, `__pycache__`, `venv`, `.venv`
-
-**Step 3: Categorize Core Nodes**
-
-Group core nodes by:
-
-- **Type**: Package vs. Working Directory
-- **Language**: JavaScript/TypeScript, Rust, Python, Go, Java, Swift, Ruby, PHP, Other
-- **Purpose**: Inferred from path and manifest (app, library, tooling, workflow, etc.)
-
-### Phase 2: User Confirmation
-
-**REQUIRED: Get user confirmation before any file creation/modification**
-
-Present summary:
-
-```
-Found {count} core nodes:
-
-Packages ({count}):
-  JavaScript/TypeScript: {count} ({list paths})
-  Rust: {count} ({list paths})
-  Python: {count} ({list paths})
-  Go: {count} ({list paths})
-  Java: {count} ({list paths})
-  Other: {count} ({list paths})
-
-Working Directories ({count}):
-  {list paths}
-
-Actions needed:
-  - Create new CLAUDE.md: {count} files
-  - Update existing CLAUDE.md: {count} files
-  - Already up-to-date: {count} files
-
-Proceed with generation/updates? (y/n/review)
-  y = Apply all
-  n = Cancel
-  review = Show what will be created/updated for each file
-```
-
-If user selects "review", show details for each file and allow selective skipping.
-
-### Phase 3: Generation
-
-## ⚠️ CRITICAL PRESERVATION RULES
-
-**THESE RULES MUST BE FOLLOWED - VIOLATIONS CAUSE DATA LOSS**
-
-### Never Modify (Tier 1: Absolute Protection)
-
-1. User-written sections (not from templates)
-2. Custom section headings (e.g., "## Our Team's Conventions")
-3. Personal notes, warnings, tips added by users
-4. Code examples written by users
-5. First-person language ("I", "we", "our approach")
-
-### Modify with Caution (Tier 2: Surgical Updates Only)
-
-1. Command lists (ADD only, never REMOVE without confirmation)
-2. Dependency lists (ADD/REMOVE based on manifest changes)
-3. File structure listings (ADD new files/dirs only)
-
-**Update Method for Tier 2:**
-
-- Use Edit tool with precise old_string/new_string
-- Never rewrite entire sections
-- Preserve formatting (maintain list style, sort order)
-
-### Can Regenerate (Tier 3: Auto-Generated)
-
-1. Sections marked: `<!-- AUTO-GENERATED - DO NOT EDIT -->`
-2. Content that exactly matches package manifest
-
-**Verification (REQUIRED before every update):**
-
-- [ ] Is section marked AUTO-GENERATED? (If no → Tier 2 rules)
-- [ ] Will change remove existing text? (If yes → flag for review)
-- [ ] Using Edit tool with specific old_string? (If no → STOP)
-- [ ] Preserving formatting? (If no → adjust)
-
-**If ANY checklist item fails → DO NOT auto-apply → Show user for confirmation**
-
-## Length Constraints
-
-**CRITICAL: All generated CLAUDE.md files MUST be concise and focused.**
-
-- **Token Limit**: 500 tokens or less (~2000 characters in English)
-- **Why**: Keeps documentation scannable and focused on essentials
-- **How to achieve**:
-  - Use bullet points instead of paragraphs
-  - Include only essential commands/dependencies
-  - Avoid verbose descriptions (5-10 words max per item)
-  - Skip redundant sections
-  - Use `[TODO]` placeholders instead of long explanations
-
-**Before writing any CLAUDE.md file:**
-
-1. Count approximate characters (use `wc -c` or similar)
-2. If >2000 characters, trim non-essential content:
-   - Remove optional sections (Structure, etc.)
-   - Shorten dependency descriptions
-   - Consolidate similar items
-3. Verify final output is ≤2000 characters
-
-## Content Template
-
-Use this adaptive template for ALL core nodes:
-
-```markdown
-# {Directory or Package Name}
-
-## Overview
-
-{1-2 sentence description of purpose}
-{For packages: "This is a {language} {type} that..."}
-{For working dirs: "This directory contains..."}
-
-## {Commands|Scripts|Tasks} (if package manifest exists)
-
-{List available commands/scripts/tasks from manifest}
-
-{Example formats by manifest type:}
-
-- package.json: `npm run {script}` or `nx {target} {project}`
-- Cargo.toml: `cargo {command}` (build, test, run, etc.)
-- pyproject.toml: `poetry run {script}` or `python -m {module}`
-- go.mod: `go {command}` (build, test, run, etc.)
-- pom.xml: `mvn {goal}` (compile, test, package, etc.)
-- build.gradle: `./gradlew {task}` or `gradle {task}`
-- Gemfile: `bundle exec {command}`
-- composer.json: `composer {script}`
-
-## {Dependencies|Requirements} (if package manifest exists)
-
-<!-- AUTO-GENERATED - Updated by /update-claude-md -->
-
-{For each main dependency:}
-
-- **{package}** ({version}) - {purpose in 5-10 words or [TODO: Add description]}
-
-## Key Files (if working directory without manifest)
-
-{For important files:}
-
-- `{filename}` - {purpose}
-
-## Structure (optional - if >5 subdirectories)
-
-{Brief overview of subdirectory organization}
-
-## Auto-Update Instructions
-
-After changes to files in this directory or subdirectories, run `/update-claude-md`
-to keep this documentation synchronized with the codebase.
-```
-
-**Template Variables:**
-
-- `{Directory or Package Name}`: Derived from path or manifest name field
-- `{language}`: Detected from manifest type (JavaScript, Rust, Python, etc.)
-- `{type}`: Detected from manifest (package, module, workspace member, etc.)
-- Commands section: Generated based on manifest type
-- Dependencies section: Only for directories with manifests
-
-## Implementation Steps
-
-**For Each Core Node:**
-
-1. **Check if CLAUDE.md exists**
-
-   ```bash
-   Read {node-path}/CLAUDE.md
-   ```
-
-2. **If file doesn't exist:**
-
-   - Select appropriate template
-   - Gather information:
-     - For packages: Read manifest for name, description, commands, dependencies
-     - For working dirs: List files, infer purpose
-   - Populate template with gathered info
-   - Add dependency descriptions (use standard descriptions for common packages, [TODO] for others)
-   - **Verify length constraint**: Check character count is ≤2000 characters
-   - If too long, trim by:
-     - Limiting dependency list to top 5-10 most important
-     - Using shorter descriptions (5-10 words max)
-     - Removing optional sections
-   - Write file using Write tool
-   - Report: "✅ Created {path}/CLAUDE.md ({n} chars)"
-
-3. **If file exists:**
-
-   - Read current content
-   - Parse into sections
-   - Compare with current project state (missing commands, new dependencies, outdated info)
-   - If significant gaps found, show user proposed updates
-   - If user confirms, apply updates using Edit tool (following preservation rules)
-   - Report: "✅ Updated {path}/CLAUDE.md ({n} sections)"
-
-4. **If file is up-to-date:**
-   - Report: "⏭️ Skipped {path}/CLAUDE.md (up-to-date)"
-
-## Success Criteria & Verification
-
-**After generation, verify (REQUIRED):**
-
-1. **File Validity**
-
-   ```bash
-   # Check file was created/updated
-   test -f {path}/CLAUDE.md
-
-   # Check file is not empty
-   test -s {path}/CLAUDE.md
-
-   # Verify contains required sections
-   grep -q "## Overview" {path}/CLAUDE.md || grep -q "## Purpose" {path}/CLAUDE.md
-   grep -q "## Auto-Update Instructions" {path}/CLAUDE.md
-   ```
-
-2. **Content Quality**
-
-   - For packages: All commands/scripts/tasks from manifest documented
-   - No placeholder text like "[Project Name]" remains
-   - All dependency entries have descriptions (even if [TODO])
-   - **File size**: Must be ≤2000 characters (verify with `wc -c`)
-   - If >2000 characters, file is TOO VERBOSE and must be trimmed
-
-3. **Required Sections Present**
-
-   - Every CLAUDE.md MUST have:
-
-     - ## Overview (or ## Purpose)
-
-     - ## Auto-Update Instructions
-
-   - Packages MUST also have:
-
-     - ## Commands (or ## Scripts or ## Tasks)
-
-     - ## Dependencies (or ## Requirements)
-
-**Mark as FAILURE if:**
-
-- ANY Write/Edit operation failed
-- File size is 0 bytes
-- File size exceeds 2000 characters (too verbose)
-- Required sections missing
-- Cannot parse as valid Markdown
-
-## Error Handling
-
-**Permission Denied**: Offer to skip, fix permissions (`chmod -R u+rw`), or cancel
-
-**Cannot Parse Manifest**: Offer to skip, create without dependency info, or cancel
-
-**No Core Nodes Found**: Display info message explaining detection criteria and suggest running on specific subdirectories
-
-For all errors: Present clear options, explain consequences, await user decision.
-
-## Summary Report Format
-
-Generate a markdown report with:
-
-- **Summary**: Totals for nodes found, files created/updated/skipped, errors
-- **Packages**: Grouped by language with status (✅/⏭️/❌)
-- **Working Directories**: List with status
-- **Errors & Warnings**: Details for any failures with recommendations
-- **Next Steps**: Checklist for user follow-up actions
-
-## Integration with /update-claude-md
-
-Every generated CLAUDE.md includes auto-update instructions. This creates a self-maintaining system:
-
-1. `/claude-init-plus` creates initial structure
-2. `/update-claude-md` keeps it current as code evolves
-3. Claude Code automatically maintains context

package/dist/content/commands/agnostic/create-pr.md

@@ -1,79 +0,0 @@
----
-name: create-pr
-description: Create or update a Graphite PR with auto-generated conventional commit messages and comprehensive descriptions based on code diffs.
----
-
-## Workflow Process
-
-This workflow uses the pr-creator agent to:
-
-1. Analyze diffs between current branch and target branch (default: main)
-2. Generate conventional commit messages if needed
-3. Create informative PR titles following conventional commits format
-4. Write comprehensive PR descriptions
-5. Create new PR or update existing PR using Graphite
-
-## Execution Steps
-
-1. **pr-creator**: Main orchestrator that will:
-   - Analyze `git diff` between current and target branch
-   - Detect the type of changes (feat, fix, refactor, etc.)
-   - If uncommitted changes exist, ASK THE USER if they would like to create a conventional commit. DO NOT commit changes without User confirmation
-   - Create PR title in format: `<type>(<scope>): <description>`
-   - Generate detailed PR description with:
-     - Summary of changes
-     - List of modified files and why
-     - Testing information
-     - Related issues
-   - Use `gt submit` or `gh pr create` to create/update PR
-   - Provide PR URL for review
-
-## Usage Examples
-
-Basic usage (create/update PR against main):
-
-```
-/create-pr
-```
-
-With custom target branch:
-
-```
-/create-pr target: develop
-/create-pr target: staging
-```
-
-With specific instructions:
-
-```
-/create-pr emphasize performance improvements
-/create-pr mark as breaking change due to API updates
-/create-pr link to issue #123
-```
-
-## Conventional Commit Types
-
-The agent will automatically detect and use appropriate types:
-
-- `feat`: New feature
-- `fix`: Bug fix
-- `docs`: Documentation changes
-- `style`: Formatting changes
-- `refactor`: Code refactoring
-- `perf`: Performance improvements
-- `test`: Test additions/changes
-- `build`: Build system changes
-- `ci`: CI configuration changes
-- `chore`: Maintenance tasks
-
-## Options
-
-The workflow supports:
-
-- Custom target branch (default: main)
-- Breaking change detection and marking
-- Issue linking
-- Stack-aware PR creation with Graphite
-- Updating existing PRs vs creating new ones
-
-Target: $ARGUMENTS

package/dist/content/commands/agnostic/daily-standup.md

@@ -1,185 +0,0 @@
----
-name: daily-standup
-description: This provides a daily standup, checking GitHub and Linear over the past day and providing a summary of what you've been working on and what you will continue to work on.
-model: claude-haiku-4-5
----
-
-## Workflow Process
-
-This workflow uses the Linear MCP and GitHub MCP to:
-
-1. Fetch all issues assigned to a user (defaults to you)
-2. Filter for incomplete/in-progress issues
-3. Organize them by priority and status
-4. Fetch GitHub activity from the last 24 hours across all repositories in the in the Uniswap GitHub organization
-5. Generate a friendly team update message combining Linear issues and GitHub activity
-
-## Execution Steps
-
-1. **Parse Arguments**: Extract optional parameters
-
-   - Linear user: defaults to "me" (authenticated user)
-   - GitHub username: must be provided by user (prompt if not specified)
-   - Parse from `$ARGUMENTS` for `user:` and `github:` parameters
-
-2. **Prompt for GitHub Username** (if not provided):
-
-   - Check if `github:` parameter was provided in `$ARGUMENTS`
-   - If not provided, ask the user: "What is your GitHub username for the Uniswap organization?"
-   - Store the response for use in GitHub queries
-
-3. **Fetch Linear Issues**: Get all issues assigned to the specified user
-
-   - Use `mcp__linear__list_issues` with assignee parameter
-   - Filter by state (exclude completed/cancelled)
-
-4. **Analyze Issues**: Review issue details including:
-
-   - Title and description
-   - Current status (Backlog, Todo, In Progress, etc.)
-   - Priority level (Urgent, High, Normal, Low)
-   - Project/team association
-   - Recent comments or updates
-   - Due dates
-
-5. **Fetch GitHub Activity** (last 24 hours):
-
-   - Use GitHub MCP search tools to find activity from the specified user
-   - Filter for Uniswap organization repositories only
-   - Include:
-     - Pull requests created, updated, or reviewed
-     - Commits pushed
-     - PR comments and reviews
-     - Issues opened or commented on
-   - Search queries to use:
-     - `org:Uniswap author:<username> created:>YYYY-MM-DD` (for PRs/issues)
-     - `org:Uniswap commenter:<username> updated:>YYYY-MM-DD` (for comments)
-     - `org:Uniswap reviewed-by:<username> updated:>YYYY-MM-DD` (for reviews)
-
-6. **Organize by Status**:
-
-   - **In Progress**: Currently working on
-   - **Todo**: Up next
-   - **Backlog**: High priority items on deck
-
-7. **Generate Update**: Create a concise message covering:
-   - What's currently being worked on (Linear issues)
-   - GitHub activity in the last 24 hours (PRs, commits, reviews)
-   - What's coming up next (Linear issues)
-   - Any blockers or help needed
-   - Format suitable for Slack/team chat
-
-## Usage Examples
-
-**Default** (your own tasks - will prompt for GitHub username):
-
-```
-/daily-standup
-```
-
-**With GitHub username specified**:
-
-```
-/daily-standup github:wkoutre
-```
-
-**Check another team member's tasks** (specify both Linear user and GitHub username):
-
-```
-/daily-standup user:john@example.com github:johndoe
-/daily-standup user:Jane Doe github:janedoe
-/daily-standup user:abc-123-user-id github:jsmith
-```
-
-**With additional instructions**:
-
-```
-/daily-standup focus on high priority items
-/daily-standup user:john@example.com brief version
-/daily-standup include estimates and blockers
-```
-
-## Parameter Format
-
-The `user:` parameter (for Linear) should be provided as:
-
-- `user:me` - Explicitly specify yourself (default)
-- `user:<email>` - e.g., `user:john@company.com`
-- `user:<name>` - e.g., `user:John Smith`
-- `user:<id>` - e.g., `user:abc-123-def-456`
-
-The `github:` parameter (for GitHub) should be provided as:
-
-- `github:<username>` - e.g., `github:wkoutre` or `github:johndoe`
-- If omitted, you will be prompted to enter your GitHub username
-
-Any text after the parameters will be treated as additional instructions for formatting or focus.
-
-## Output Format
-
-The generated update will include:
-
-**GitHub Activity (Last 24 Hours)**:
-
-- PRs opened, updated, or merged
-- Code reviews completed
-- Commits pushed (with repo names)
-- Comments on PRs/issues
-- Links to relevant PRs and commits
-
-**Currently Working On** (In Progress):
-
-- Issue titles with priority indicators
-- Links to issues
-- Brief context
-
-**Up Next** (Todo):
-
-- Prioritized list of upcoming work
-- Due dates if relevant
-
-**On Deck** (Backlog - High Priority):
-
-- Important items queued up
-
-**Blockers/Help Needed**:
-
-- Issues that may need team support
-- Questions or dependencies
-
-**Notes**:
-
-- Recent activity or comments
-- Additional context
-
-The message will be conversational and ready to paste into Slack or share with your team.
-
-## Implementation Details
-
-The workflow will:
-
-- Parse `$ARGUMENTS` to extract `user:` and `github:` parameters
-  - Linear user: default to "me" if not specified
-  - GitHub username: if not found in arguments, prompt user with: "What is your GitHub username for the Uniswap organization?"
-- **IMPORTANT**: Always prompt for GitHub username FIRST before fetching any data if it's not provided in arguments
-- Calculate the timestamp for 24 hours ago from current time (format: YYYY-MM-DD)
-- **Fetch GitHub Activity**:
-  - Use the provided or prompted GitHub username
-  - Use `mcp__github__search_issues` with queries like:
-    - `org:Uniswap author:<username> created:>YYYY-MM-DD` (PRs/issues created)
-    - `org:Uniswap involves:<username> updated:>YYYY-MM-DD` (all activity including comments, reviews)
-  - Get PR details for any PRs found using `mcp__github__get_pull_request`
-  - Include PR status (open/merged/closed) and review status
-- **Fetch Linear Issues**:
-  - Call Linear MCP with the assignee parameter
-  - Prioritize urgent/high priority items
-  - Note any issues with recent activity or comments
-- **Combine and Format**:
-  - Start with GitHub activity section (most recent work)
-  - Follow with Linear issues organized by status
-  - Identify potential blockers based on status, priority, and age
-  - Suggest what help might be needed from the team
-- Keep the tone professional but conversational
-- Handle cases where no incomplete tasks or GitHub activity exist gracefully
-
-Arguments: $ARGUMENTS