@uniswap/ai-toolkit-nx-claude 0.5.12 → 0.5.14

package/dist/content/agents/agnostic/claude-docs-initializer.md

@@ -246,7 +246,7 @@ Every CLAUDE.md file MUST start with a timestamp header as the very first line:
 <!-- CUSTOM:END -->
 ```
 
-#### For Package/Module CLAUDE.md:
+#### For Package/Module CLAUDE.md
 
 ```markdown
 > **Last Updated:** YYYY-MM-DD
@@ -286,7 +286,7 @@ Every CLAUDE.md file MUST start with a timestamp header as the very first line:
 <!-- CUSTOM:END -->
 ```
 
-#### For Feature/Component CLAUDE.md:
+#### For Feature/Component CLAUDE.md
 
 ```markdown
 # CLAUDE.md - [Feature/Component Name]

package/dist/content/agents/agnostic/claude-docs-manager.md

@@ -207,7 +207,7 @@ verified_facts:
 
 For each affected CLAUDE.md file, determine content updates:
 
-#### When to Create New CLAUDE.md Files:
+#### When to Create New CLAUDE.md Files
 
 **Create a NEW CLAUDE.md when**:
 
@@ -227,7 +227,7 @@ For each affected CLAUDE.md file, determine content updates:
 - Pure type definition directories
 - Asset/static file directories
 
-#### If Root CLAUDE.md:
+#### If Root CLAUDE.md
 
 Include:
 
@@ -239,7 +239,7 @@ Include:
 - Repository structure overview (immediate subdirectories only)
 - Code quality standards and setup
 
-#### If Package/Module CLAUDE.md:
+#### If Package/Module CLAUDE.md
 
 Include:
 
@@ -251,7 +251,7 @@ Include:
 - Usage patterns and examples
 - Package-specific guidelines
 
-#### If Feature/Component CLAUDE.md:
+#### If Feature/Component CLAUDE.md
 
 Include:
 
@@ -266,7 +266,7 @@ Include:
 
 Process all identified CLAUDE.md files:
 
-#### For Each Affected File:
+#### For Each Affected File
 
 1. **New CLAUDE.md**: Generate complete structure appropriate for that level
 2. **Existing CLAUDE.md**:
@@ -510,7 +510,7 @@ error: # Only if success: false
 
 ## Directory Importance Examples
 
-### ✅ **Deserves CLAUDE.md**:
+### ✅ **Deserves CLAUDE.md**
 
 ```
 /packages/auth              # Package with package.json
@@ -521,7 +521,7 @@ error: # Only if success: false
 /packages/ui/components     # Component library root
 ```
 
-### ❌ **Skip CLAUDE.md**:
+### ❌ **Skip CLAUDE.md**
 
 ```
 /src/utils                  # Just utility functions

package/dist/content/agents/agnostic/context-loader.md

@@ -120,6 +120,7 @@ Return a structured report containing:
 ### Core Context Report
 
 - `summary`: Multi-level executive summary
+
   ```
   {
     executive: string,          // 2-3 sentences
@@ -127,7 +128,9 @@ Return a structured report containing:
     technical: string           // Full technical overview
   }
   ```
+
 - `relevance_scores`: Context relevance ratings
+
   ```
   {
     overall: number,           // 0-100
@@ -135,7 +138,9 @@ Return a structured report containing:
     by_pattern: Record<string, number>
   }
   ```
+
 - `key-components`: Core files/modules with metadata
+
   ```
   {
     path: string,
@@ -146,7 +151,9 @@ Return a structured report containing:
     summary: string
   }[]
   ```
+
 - `patterns`: Identified patterns with advanced metrics
+
   ```
   {
     pattern: string,
@@ -158,7 +165,9 @@ Return a structured report containing:
     impact: 'low' | 'medium' | 'high'
   }[]
   ```
+
 - `dependencies`: External dependencies with risk assessment
+
   ```
   {
     type: 'library' | 'service' | 'config',
@@ -169,7 +178,9 @@ Return a structured report containing:
     alternatives?: string[]
   }[]
   ```
+
 - `data-flow`: Hierarchical data flow representation
+
   ```
   {
     overview: string,
@@ -182,7 +193,9 @@ Return a structured report containing:
     }[]
   }
   ```
+
 - `insights`: Key discoveries and recommendations
+
   ```
   {
     type: 'pattern' | 'risk' | 'opportunity' | 'gotcha',

package/dist/content/agents/agnostic/debug-assistant.md

@@ -26,6 +26,7 @@ You are **debug-assistant**, an advanced debugging specialist focused on compreh
 ### Root Cause Analysis
 
 - **rootCauses**: Ranked hypotheses with confidence scores
+
   ```
   [
     {
@@ -48,6 +49,7 @@ You are **debug-assistant**, an advanced debugging specialist focused on compreh
 ### Fix Strategy
 
 - **fixPlan**: Detailed remediation steps
+
   ```
   {
     "immediate": [
@@ -70,6 +72,7 @@ You are **debug-assistant**, an advanced debugging specialist focused on compreh
 ### Code Patches
 
 - **patches**: Concrete code changes with context
+
   ```
   [
     {
@@ -85,6 +88,7 @@ You are **debug-assistant**, an advanced debugging specialist focused on compreh
 ### Test Coverage
 
 - **tests**: Regression and validation tests
+
   ```
   [
     {
@@ -100,6 +104,7 @@ You are **debug-assistant**, an advanced debugging specialist focused on compreh
 ### Prevention Recommendations
 
 - **prevention**: Long-term improvements
+
   ```
   {
     "codeChanges": [

package/dist/content/agents/agnostic/plan-reviewer.md

@@ -27,6 +27,7 @@ Critically analyze implementation plans WITHOUT writing any code. Focus on revie
 
 **MANDATORY DEEP THINKING PHASE:**
 Before providing any review, you MUST:
+
 1. Deeply read and understand the entire plan
 2. **Integrate context_findings if provided** - Use the deep understanding from context-loader
 3. Consider multiple potential issues with the plan
@@ -36,6 +37,7 @@ Before providing any review, you MUST:
 7. Assess conciseness vs over-documentation
 
 **Review Steps:**
+
 1. **Context Integration**: If context_findings provided, use them as foundation for review:
    - Cross-reference plan against key_components identified by context-loader
    - Verify plan follows the patterns and conventions already discovered
@@ -60,7 +62,7 @@ summary: |
 strengths:
   - [What the plan does well - be specific]
   - [Areas where plan shows good understanding]
-  
+
 concerns:
   - issue: [Specific concern about the plan]
     severity: low|medium|high|critical
@@ -68,16 +70,22 @@ concerns:
     suggestion: [How to address it in the plan]
 
 gaps:
-  - missing: [What's missing from the plan that's critical for implementation. Note: Do NOT flag missing testing plans, success criteria, risk matrices, or agent assignments as gaps - these are intentionally omitted]
+  - missing:
+      [
+        What's missing from the plan that's critical for implementation. Note: Do NOT flag missing testing plans,
+        success criteria,
+        risk matrices,
+        or agent assignments as gaps - these are intentionally omitted,
+      ]
     rationale: [Why this gap matters for the stated goal]
     suggestion: [How to fill this gap]
-    
+
 improvements:
   - area: [What could be better in the plan]
     current: [Current approach described in plan]
     suggested: [Better approach]
     rationale: [Why the suggested approach is better]
-    
+
 feasibility-assessment:
   complexity: low|medium|high
   risks:
@@ -97,6 +105,7 @@ scope-validation:
 ## Guidelines
 
 **ABSOLUTE REQUIREMENTS:**
+
 1. **NO CODE WRITING** - Do NOT write any implementation code, only review plans
 2. **NO EXTRA SUGGESTIONS** - Do NOT suggest features not in the original plan:
    - NO backwards compatibility suggestions unless plan requested it
@@ -110,6 +119,7 @@ scope-validation:
 7. **VALIDATE CONCISENESS** - Plans should be strategic roadmaps, not exhaustive documentation
 
 **Review Principles:**
+
 - **Leverage context_findings when available** - Don't duplicate analysis already done by context-loader
 - Focus on plan quality, not implementation details
 - Verify plan follows existing architectural decisions (especially those identified in context_findings)
@@ -118,12 +128,14 @@ scope-validation:
 - Be explicit about scope adherence
 
 **Context Integration Best Practices:**
+
 - If context_findings are provided, treat them as authoritative
 - Verify plan builds upon the patterns and conventions already identified
 - Flag if plan contradicts the gotchas and edge cases discovered
 - Reference specific files from key_components when assessing plan accuracy
 
 **Quality Focus Areas:**
+
 - Is the plan actionable without ambiguity?
 - Are all steps concrete and specific?
 - Have edge cases been considered (including those from context)?
@@ -133,12 +145,14 @@ scope-validation:
 - Are there any missing critical steps or considerations?
 
 **Critical Scope Enforcement:**
+
 - Flag any suggestions for backwards compatibility not explicitly requested
 - Identify any legacy support not specifically required
 - Point out future-proofing or nice-to-haves beyond current needs
 - Ensure plan implements EXACTLY what was asked for
 
 **What NOT to Flag as Problems:**
+
 - Missing testing plans (testing is handled during execution, not planning)
 - Missing success criteria checklists (implementer validates)
 - Missing comprehensive risk matrices (only critical risks should be documented)
@@ -148,6 +162,7 @@ scope-validation:
 - Plans being "too concise" if they cover all critical information
 
 **What SHOULD be Flagged:**
+
 - Over-documentation or exhaustive details that make plan hard to use
 - Missing critical implementation steps or decisions
 - Unclear API interfaces when needed
@@ -155,4 +170,4 @@ scope-validation:
 - Scope creep or extras not requested
 - Plans where strategic direction is unclear
 
-Remember: Your role is critical analysis of the plan's quality and feasibility. When context_findings are provided, you're building on deep reconnaissance already performed. Focus on ensuring the plan is complete, accurate, concise, and implementable without unnecessary additions.
+Remember: Your role is critical analysis of the plan's quality and feasibility. When context_findings are provided, you're building on deep reconnaissance already performed. Focus on ensuring the plan is complete, accurate, concise, and implementable without unnecessary additions.

package/dist/content/agents/agnostic/researcher.md

@@ -70,6 +70,7 @@ You are **researcher**, a specialized research subagent that synthesizes externa
 #### C. Technology Comparison
 
 - **Evaluation Framework**:
+
   ```
   Performance: Benchmarks, latency, throughput, resource usage
   Scalability: Horizontal/vertical scaling capabilities, limits
@@ -80,6 +81,7 @@ You are **researcher**, a specialized research subagent that synthesizes externa
   Cost: Licensing, infrastructure requirements, operational costs
   Security: Known vulnerabilities, security features, compliance
   ```
+
 - **Comparison Matrix**: Create structured comparisons with weighted criteria
 - **Use Case Alignment**: Match technology characteristics to specific requirements
 - **Sources**: Official benchmarks, case studies, Stack Overflow surveys, GitHub statistics

package/dist/content/commands/agnostic/address-pr-issues.md

@@ -51,7 +51,7 @@ Review the specified pull request, address all review comments, and fix any CI/C
 
    - If owner/repo not provided, infer from current git repository
    - Validate that the PR exists and is accessible
-   - If at any point it becomes clear the user has not installed the GitHub MCP, direct them to the following Notion page where they can follow a tutorial for how to correctly install the GitHub MCP on Claude Code: https://www.notion.so/uniswaplabs/Using-a-GitHub-MCP-with-Claude-Code-270c52b2548b8015b11ee5e905796cb5
+   - If at any point it becomes clear the user has not installed the GitHub MCP, direct them to the following Notion page where they can follow a tutorial for how to correctly install the GitHub MCP on Claude Code: <https://www.notion.so/uniswaplabs/Using-a-GitHub-MCP-with-Claude-Code-270c52b2548b8015b11ee5e905796cb5>
 
 2. **Analyze PR**:
 

package/dist/content/commands/agnostic/claude-docs.md

@@ -13,6 +13,7 @@ Intelligently manages CLAUDE.md documentation files across your repository. Supp
 Accept natural language input and extract intent:
 
 - **Input text**: Parse the user's request to understand intent
+
   - **"init" keywords**: "init", "initialize", "setup", "create", "bootstrap", "start"
     - Triggers: "init docs", "initialize documentation", "setup CLAUDE.md files"
     - Routes to: **INIT Path** (claude-docs-initializer agent)
@@ -24,6 +25,7 @@ Accept natural language input and extract intent:
     - Path references: "in packages/ui", "for the UI package", specific paths mentioned
 
 - **Mode determination**:
+
   1. Explicitly check for init keywords first → Use **INIT Path**
   2. Otherwise → Use **UPDATE Path** (default)
   3. Extract any file limit specification (e.g., "init with max 50 files per agent")
@@ -45,6 +47,7 @@ Accept natural language input and extract intent:
 ### Task for Update
 
 1. **Analyze Change Significance**:
+
    - Identify significant vs insignificant changes
    - Filter out minor fixes, typos, formatting
 
@@ -55,6 +58,7 @@ Accept natural language input and extract intent:
 ### Delegation for Update
 
 Invoke **claude-docs-manager** with:
+
 - `changes`: Complete list of all significant file changes
   - `filePath`: Absolute path to changed file
   - `changeType`: added|modified|deleted
@@ -68,6 +72,7 @@ Invoke **claude-docs-manager** with:
 - `rootGuidelines`: (Optional) Key guidelines from existing root CLAUDE.md
 
 The agent will:
+
 - Identify ALL CLAUDE.md files that need updates
 - Update each file with appropriate scope (component/module/package/root)
 - Create new CLAUDE.md files where needed
@@ -76,6 +81,7 @@ The agent will:
 ### Output for Update
 
 Return results from claude-docs-manager:
+
 ```yaml
 summary: |
   Successfully updated N CLAUDE.md files
@@ -148,12 +154,14 @@ Please be patient - the process is working even during apparent pauses.
 ### Task for Init
 
 1. **Repository Analysis**:
+
    - Use git to quickly understand structure: `git ls-files | wc -l`
    - Find major boundaries: packages, apps, services
    - Count files per area: `git ls-files [area] | wc -l`
    - Assess complexity by file types and patterns
 
 2. **Intelligent Area Splitting**:
+
    - Default: 100 files at most per agent (or user-specified limit)
    - Split areas exceeding limits into logical sub-areas
    - Identify natural boundaries (pages vs components, routes vs services)
@@ -165,6 +173,7 @@ Please be patient - the process is working even during apparent pauses.
    - Track dependencies between levels
 
 Example execution plan for large monorepo:
+
 ```
 Level 1 (Parallel leaf documentation):
   - Agent A: "Document user-facing frontend pages in /frontend/pages/user (50 files)"
@@ -195,11 +204,13 @@ Execute hierarchical parallelized documentation creation:
 #### Phase 1: Repository Analysis
 
 **⚠️ CRITICAL: ALWAYS prefer git commands over find/glob for discovery!**
+
 - Git automatically excludes node_modules, build outputs, and ignored files
 - Only use find/glob as fallback for non-git repositories
 - If using find, MUST use `*/node_modules/*` not `./node_modules/*` for exclusions
 
 Quickly analyze repository structure using git:
+
 - Get total file count: `git ls-files | wc -l`
 - Find package boundaries: `git ls-files | grep 'package\.json$'`
 - Identify tech stacks used and major directories and their sizes
@@ -214,6 +225,7 @@ Based on analysis, determine splitting strategy
 Invoke multiple **claude-docs-initializer** agents in PARALLEL:
 
 For each Level 1 agent:
+
 - `target`: "Document [specific area description] in [path]"
 - `siblingContext`: "Other agents are documenting: [list of other Level 1 areas]"
 
@@ -243,6 +255,7 @@ After all Level 2 agents complete, invoke **claude-docs-manager**:
 ### Output for Init
 
 Return aggregated results from all agents (claude-docs-initializer for all levels except repository root, claude-docs-manager for repository root only):
+
 ```yaml
 summary: |
   Repository analysis: [monorepo with X packages | single app | library]
@@ -254,41 +267,41 @@ executionPlan:
   level1Agents: 5 # Parallel leaf documentation
   level2Agents: 2 # Area roots
   level3Agents: 1 # Repository root
-  totalExecutionTime: "2m 34s"
+  totalExecutionTime: '2m 34s'
 
 createdFilesByLevel:
   leafDocumentation:
-    - agent: "frontend-user-pages"
+    - agent: 'frontend-user-pages'
       files: 3
-      paths: ["/frontend/pages/user/CLAUDE.md", ...]
-    - agent: "frontend-admin-pages"
+      paths: ['/frontend/pages/user/CLAUDE.md', ...]
+    - agent: 'frontend-admin-pages'
       files: 2
-      paths: ["/frontend/pages/admin/CLAUDE.md", ...]
+      paths: ['/frontend/pages/admin/CLAUDE.md', ...]
 
   areaRoots:
-    - path: "/frontend/CLAUDE.md"
-      synthesizedFrom: ["frontend-user-pages", "frontend-admin-pages", "frontend-components"]
-    - path: "/backend/CLAUDE.md"
-      synthesizedFrom: ["backend-routes", "backend-services"]
+    - path: '/frontend/CLAUDE.md'
+      synthesizedFrom: ['frontend-user-pages', 'frontend-admin-pages', 'frontend-components']
+    - path: '/backend/CLAUDE.md'
+      synthesizedFrom: ['backend-routes', 'backend-services']
 
   repositoryRoot:
-    - path: "/CLAUDE.md"
-      synthesizedFrom: ["frontend-root", "backend-root", "database"]
+    - path: '/CLAUDE.md'
+      synthesizedFrom: ['frontend-root', 'backend-root', 'database']
 
 architecturalFindings: # Aggregated from all agents
-  frontend: "Next.js 14 with App Router, Tailwind CSS, 234 total components"
-  backend: "Express with layered architecture, 23 RESTful endpoints"
-  database: "PostgreSQL with Prisma ORM, 15 models"
-  patterns: "Consistent use of TypeScript, feature-based organization"
+  frontend: 'Next.js 14 with App Router, Tailwind CSS, 234 total components'
+  backend: 'Express with layered architecture, 23 RESTful endpoints'
+  database: 'PostgreSQL with Prisma ORM, 15 models'
+  patterns: 'Consistent use of TypeScript, feature-based organization'
 
 recommendations: # Collected from all agents
-  - "Split large UserDashboard component (500+ lines)"
-  - "Add CLAUDE.md for growing analytics module"
-  - "Consider documenting complex auth flow separately"
+  - 'Split large UserDashboard component (500+ lines)'
+  - 'Add CLAUDE.md for growing analytics module'
+  - 'Consider documenting complex auth flow separately'
 
 errors: # Any failures across all agents
-  - agent: "frontend-utils"
-    error: "Failed to analyze due to circular dependencies"
+  - agent: 'frontend-utils'
+    error: 'Failed to analyze due to circular dependencies'
 ```
 
 ### Examples for Init
@@ -320,6 +333,7 @@ errors: # Any failures across all agents
 When user runs: `/claude-docs init`
 
 1. **Quick Analysis** determines:
+
    ```
    Total files: 2500
    Frontend: /app (1800 files)
@@ -328,6 +342,7 @@ When user runs: `/claude-docs init`
    ```
 
 2. **Intelligent Splitting** creates plan:
+
    ```
    Frontend needs splitting (1800 > 100):
    - /app/(user) → 450 files (complex pages) → split to 150x3
@@ -343,6 +358,7 @@ When user runs: `/claude-docs init`
    ```
 
 3. **Hierarchical Execution**:
+
    ```
    Level 1 (14 agents in parallel):
    ├── frontend-user-1 (150 files)
@@ -390,4 +406,4 @@ When user runs: `/claude-docs init`
 5. **Parallel Execution**: Use Task tool to spawn multiple agents simultaneously
 6. **Session-First Default**: For updates, prioritize session changes unless specified
 7. **Preserve Custom Content**: Both agents respect user-added content sections
-8. **Documentation Hierarchy**: Agents create docs only at their assigned level
+8. **Documentation Hierarchy**: Agents create docs only at their assigned level

package/dist/content/commands/agnostic/gen-tests.md

@@ -58,7 +58,7 @@ For complex test generation (multiple files or --type all), use orchestration:
 
 ## Delegation
 
-### Simple Case (single file, unit tests):
+### Simple Case (single file, unit tests)
 
 Invoke **test-writer** with:
 
@@ -69,7 +69,7 @@ Invoke **test-writer** with:
 - `requirements`: from `--requirements` if provided
 - `context`: dependency information if needed
 
-### Complex Case (multiple files or integration/e2e):
+### Complex Case (multiple files or integration/e2e)
 
 Invoke **agent-orchestrator** to coordinate:
 

package/dist/content/commands/agnostic/research.md

@@ -46,4 +46,4 @@ Return structured research findings:
 - `codebase_insights`: Relevant patterns from code (when applicable)
 - `recommendations`: Actionable next steps
 - `warnings`: Important gotchas or risks
-- `references`: Sources consulted with links
+- `references`: Sources consulted with links