/plan

Usage

/plan 001                    # Create plan for issue 001 (auto-detect project)
/plan 001 --project coordinatr  # Explicit project
/plan SPIKE-003              # Plan for a spike (creates exploration plans)
/plan 003 --second-opinion   # Get peer review from Gemini CLI before finalizing

Issue Type Detection

Execution Flow

For Task/Bug

1. Load Context:
   bash

   Read: ideas/[project]/issues/###-*/TASK.md (or BUG.md)
   Read: spaces/[project]/docs/specs/*.md (if implements: field exists)
   Glob: spaces/[project]/docs/adrs/ADR-*.md
   Glob: resources/research/*.md

   If the issue has an

   implements:

   field, load that specific spec section:
   bash

   Read: spaces/[project]/docs/specs/required-features.md  # Extract relevant section

2. Cross-Project Pattern Search:
   bash

   # Search other projects for similar implementations
   Grep: spaces/*/src/ for relevant patterns

   Include relevant references in plan:
   markdown

   ## Related Implementations
   
   Found similar patterns in other projects:
   - `spaces/yourbench/src/auth/clerk.ts` - Clerk auth setup
   - `spaces/coordinatr/src/lib/session.ts` - Session handling

3. Library Documentation (automatic for integrations):
   MANDATORY when task involves:
   • Installing/configuring external libraries or SDKs
   • Framework integrations (auth providers, databases, APIs)
   • Third-party services (Clerk, Stripe, AWS services, etc.)
   Process:
   bash

   # 1. Resolve library ID
   mcp__context7__resolve-library-id: {libraryName}
   
   # 2. Fetch current documentation
   mcp__context7__query-docs: {context7CompatibleLibraryID}
   
   # 3. Search for recent patterns/best practices
   WebSearch: "{library} {framework} integration 2026"

   Document findings:
   markdown

   ## Library Documentation Validation
   
   **{Library Name}** (validated YYYY-MM-DD):
   - Current version: X.Y.Z
   - Key integration patterns: [summary]
   - Recommended approach: [based on current docs]

4. Generate Plan:
   • Break work into logical phases
   • Each phase has clear deliverables
   • Include checkpoints between phases
   • Add "Done When" criteria
5. Write PLAN.md:
   • Present phases, estimated effort, dependencies
   • Include "Library Documentation Validation" section (if applicable)
   • Include "Second Opinion Analysis" section (if requested)
6. Commit Suggestion:
   • Ask: "Commit PLAN.md to ideas repo? (yes/no)"

For Spike (Exploration)

1. Load Context: Read SPIKE.md (questions, success criteria, time box)
2. Gather Approaches:
   • Ask: "How many approaches to explore?" → N
   • For each: "Describe approach N?"
3. Generate: Create PLAN-N.md for each approach
4. Commit Suggestion: Ask to commit all plan files

Task Plan Example

# Implementation Plan: 001 Research Auth Patterns

## Overview
Research authentication patterns for Coordinatr's multi-tenant architecture.

## Phase 1: Survey Existing Solutions

### 1.1 - Research Auth Libraries
- [ ] Review Better Auth documentation
- [ ] Compare with Auth.js and Lucia
- [ ] Document trade-offs

### 1.2 - Multi-Tenant Patterns
- [ ] Research team-based auth patterns
- [ ] Review how Slack, Linear handle it
- [ ] [CHECKPOINT] Summary document complete

## Phase 2: Architecture Proposal

### 2.1 - Draft Architecture
- [ ] Create architecture diagram
- [ ] Document token strategy
- [ ] Define permission model

### 2.2 - Review
- [ ] Self-critique against requirements
- [ ] [CHECKPOINT] Architecture doc complete

## Done When
- [ ] Auth library recommendation documented
- [ ] Architecture proposal in ideas/coordinatr/docs/
- [ ] Trade-offs and risks identified

Second Opinion Feature

--second-opinion

1. Send plan to Gemini CLI for review
2. Validate each recommendation against Context7 docs
3. Claude makes final decision on each:
   • ✅ ACCEPT: Recommendation validated AND improves plan
   • ⚠️ MODIFY: Good idea but needs adjustment
   • ❌ REJECT: Invalid or not applicable
4. Document all decisions in "Second Opinion Analysis" section

• Gemini CLI installed and functional

• --second-opinion

  flag explicitly passed

Workflow

/spec → /issue → /plan {ID} → (work phases) → /complete {ID}
                    ↓
          Load spec section from implements: field

• Task/Bug:

  ideas/{project}/issues/###-*/PLAN.md

• Spike:

  ideas/{project}/issues/###-*/PLAN-1.md

  ,

  PLAN-2.md

  , etc.