create-plan

Implementation Plan

You are tasked with creating detailed implementation plans through an interactive, iterative process. You should be skeptical, thorough, and work collaboratively with the user to produce high-quality technical specifications.

Replace PROJ in ticket references with your Linear team's prefix from .catalyst/config.json.

Prerequisites

# Check project setup (thoughts, CLAUDE.md snippet, config)
if [[ -f "${CLAUDE_PLUGIN_ROOT}/scripts/check-project-setup.sh" ]]; then
  "${CLAUDE_PLUGIN_ROOT}/scripts/check-project-setup.sh" || exit 1
fi

# Auto-discover most recent research (workflow context + filesystem fallback)
RECENT_RESEARCH=""
if [[ -f "${CLAUDE_PLUGIN_ROOT}/scripts/workflow-context.sh" ]]; then
  RECENT_RESEARCH=$("${CLAUDE_PLUGIN_ROOT}/scripts/workflow-context.sh" recent research)
fi
if [[ -n "$RECENT_RESEARCH" ]]; then
  echo "📋 Auto-discovered recent research: $RECENT_RESEARCH"
else
  echo "⚠️ No recent research found in workflow context or filesystem"
fi

Session Tracking

SESSION_SCRIPT="${CLAUDE_PLUGIN_ROOT}/scripts/catalyst-session.sh"
if [[ -x "$SESSION_SCRIPT" ]]; then
  CATALYST_SESSION_ID=$("$SESSION_SCRIPT" start --skill "create-plan" \
    --ticket "${TICKET_ID:-}" \
    --workflow "${CATALYST_SESSION_ID:-}")
  export CATALYST_SESSION_ID
  "$SESSION_SCRIPT" phase "$CATALYST_SESSION_ID" "planning" --phase 1
fi

Initial Response

Auto-discovery has already run in Prerequisites above. Check its output and follow this priority:

1. If user provided parameters (file path or ticket reference):
2. Use the provided path (user override)
3. Read any provided files FULLY
4. If Prerequisites also discovered research (📋), mention it and ask if it should inform the plan
5. Begin the research process

1. If no parameters provided AND Prerequisites discovered research (📋):
2. Show the discovered research path
3. Ask if it should be used as context for the plan
4. Wait for user's confirmation

1. If no parameters AND no research found (⚠️):
2. Ask for: task/ticket description, context/constraints, related research
3. Wait for user's input

Process Steps

Step 1: Context Gathering & Initial Analysis

1. Read all mentioned files immediately and FULLY:
2. Ticket files, research documents, related plans, JSON/data files
3. IMPORTANT: Use the Read tool WITHOUT limit/offset parameters
4. CRITICAL: Read these files yourself before spawning sub-tasks

1. Extract ticket and update Linear state:

If a ticket is detected (from the research document's source_ticket frontmatter, from the command argument, or from context), update ticket status to stateMap.planning from config using Linearis CLI (run linearis issues usage for syntax). If Linearis CLI is not available, skip silently and continue planning.

1. Gather context using research sub-agents — use the same agent palette and orientation

process as /catalyst-dev:research-codebase (that skill is the single source of truth for how codebase research works). For planning, focus agents on the specific ticket/task scope rather than broad exploration:

• codebase-locator — find all files related to the ticket/task
• codebase-analyzer — understand how the current implementation works
• thoughts-locator — find existing thoughts documents about this feature (if relevant)

1. Read all files identified by research tasks FULLY into the main context

1. Analyze and verify understanding:
2. Cross-reference ticket requirements with actual code
3. Identify discrepancies, assumptions, and true scope

1. Present informed understanding and focused questions:
2. Show what you found with file:line references
3. Only ask questions you genuinely cannot answer through code investigation

Step 2: Research & Discovery

After getting initial clarifications:

1. If the user corrects any misunderstanding:
2. Spawn new research tasks to verify — don't just accept corrections
3. Only proceed once you've verified the facts yourself

1. Create a research todo list using TodoWrite

1. Spawn parallel sub-tasks for comprehensive research:

For local codebase:

• codebase-locator — find specific files
• codebase-analyzer — understand implementation details
• codebase-pattern-finder — find similar features to model after

For external research:

• external-research — framework patterns and best practices from popular repos

For historical context:

• thoughts-locator / thoughts-analyzer — find past research, plans, decisions

1. Wait for ALL sub-tasks to complete before proceeding

1. Present findings and design options with pros/cons for each approach

Step 3: Plan Structure Development

Once aligned on approach:

1. Create initial plan outline showing phases and what each accomplishes
2. Get feedback on structure before writing details

Step 4: Detailed Plan Writing

After structure approval:

1. Gather metadata:

``bash CURRENT_ISO_DATETIME=$(date -Iseconds) CURRENT_DATE=$(date +%Y-%m-%d) GIT_COMMIT_SHORT=$(git rev-parse --short HEAD) GIT_BRANCH=$(git branch --show-current) REPO_NAME=$(basename "$(git rev-parse --show-toplevel)") ``

IMPORTANT: Document Storage Rules

• ALWAYS write to thoughts/shared/plans/
• NEVER write to thoughts/searchable/ (read-only search index)

1. Write the plan to thoughts/shared/plans/YYYY-MM-DD-PROJ-XXXX-description.md
2. With ticket: 2025-01-08-PROJ-123-parent-child-tracking.md
3. Without ticket: 2025-01-08-improve-error-handling.md

1. Use this template structure (frontmatter comes BEFORE the heading):

---
date: { CURRENT_ISO_DATETIME }
researcher: claude
git_commit: { GIT_COMMIT_SHORT }
branch: { GIT_BRANCH }
repository: { REPO_NAME }
topic: "{PLAN_TITLE}"
tags: [plan, implementation, { RELEVANT_COMPONENT_TAGS }]
status: ready_for_implementation
last_updated: { CURRENT_DATE }
last_updated_by: claude
type: implementation_plan
source_ticket: { TICKET-ID or null }
source_research: "[[research-doc-filename]]" # or null
---

# [Feature/Task Name] Implementation Plan

## Overview

[Brief description of what we're implementing and why]

## Current State Analysis

[What exists now, what's missing, key constraints discovered]

## Desired End State

[Specification of the desired end state and how to verify it]

### Key Discoveries:

- [Important finding with file:line reference]
- [Pattern to follow]
- [Constraint to work within]

## What We're NOT Doing

[Explicitly list out-of-scope items to prevent scope creep]

## Implementation Approach

[High-level strategy and reasoning]

## Phase 1: [Descriptive Name]

### Overview

[What this phase accomplishes]

### Tests First (Red):

Define the expected behavior before writing implementation code.

#### 1. [Test File/Group]

**File**: `tests/path/to/feature.test.ext` **Tests to write**:

// Test describing expected behavior — should FAIL before implementation

### Implementation (Green):

Write the minimum code to make the tests pass.

#### 1. [Component/File Group]

**File**: `path/to/file.ext` **Changes**: [Summary of changes]

// Specific code to add/modify

### Refactor (if needed):

[Any cleanup, extraction, or simplification to do while tests stay green]

### Success Criteria:

#### Automated Verification:

- [ ] Unit tests pass: `make test`
- [ ] Type checking passes: `make check`
- [ ] Linting passes: `make lint`

#### Manual Verification:

- [ ] Feature works as expected when tested
- [ ] No regressions in related features

---

## Phase 2: [Descriptive Name]

[Similar structure — always Tests First → Implementation → Refactor]

---

## Testing Strategy (TDD)

**Approach: Test-Driven Development (Red → Green → Refactor)**

Each phase writes tests BEFORE implementation