메뉴
Create detailed implementation plans through interactive research and iteration. Use when starting a new feature or ticket that needs a comprehensive technical plan. Saves plans to the repo-appropriate plans location and emits Personal Knowledge events when applicable.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | create-plan |
| description | Create detailed implementation plans through interactive research and iteration. Use when starting a new feature or ticket that needs a comprehensive technical plan. Saves plans to the repo-appropriate plans location and emits Personal Knowledge events when applicable. |
| allowed-tools | Read, Write, Bash, Agent, mcp__plugin_claude-mem_mcp-search__smart_search, mcp__plugin_claude-mem_mcp-search__smart_outline, mcp__plugin_claude-mem_mcp-search__smart_unfold |
Create detailed implementation plans through an interactive, iterative process.
For code file exploration, prefer smart-explore tools (4-18x token savings):
| Tool | Use instead of |
|---|---|
smart_search(query, path) — discover files + symbols in one call | Glob + Grep discovery loop |
smart_outline(file_path) — structural skeleton (~1-2k tokens) | Read on files > 100 lines |
smart_unfold(file_path, symbol_name) — one symbol's full source | Read on large files for one function |
Rule: Use Read only for files < 100 lines or non-code files (markdown, JSON, config, tickets).
If no parameters: ask for task/ticket description and wait. If file path provided: read it FULLY and begin immediately.
Tip: /create-plan thoughts/shared/tickets/eng_1234.md
Read all mentioned files fully (no limit/offset). For code files > 100 lines, prefer smart_outline + smart_unfold. Do NOT spawn sub-tasks before reading.
Before spawning sub-agents, pull past work on this feature/ticket.
mcp__plugin_claude-mem_mcp-search__search(query="<feature/ticket topic>", project="<repo>", limit=15, orderBy="relevance")mcp__plugin_claude-mem_mcp-search__get_observations(ids=[...])Embed results into sub-agent prompts and include a brief summary in "Present Understanding" so the user sees what past work informs this plan. Skip silently if unavailable.
Spawn parallel agents:
Wait for ALL to complete. Then read all identified relevant files.
Based on [ticket] and codebase research, I understand we need to [summary].
Found:
- [Current implementation at file:line]
- [Pattern to follow]
Questions (only what code can't answer):
- [Question 1]
Propose phasing first, get buy-in before writing details.
Path:
plans/YYYY-MM-DD-description.mdthoughts/shared/plans/YYYY-MM-DD[-ENG-XXXX]-description.mdplans/, docs/plans/, or thoughts/shared/plans/ convention.# [Feature] Implementation Plan
## Overview
## Current State Analysis
## Desired End State
### Key Discoveries
## What We're NOT Doing
## Implementation Approach
## Phase 1: [Name]
### Overview
### Changes Required
#### 1. [Component]
**File**: `path/to/file`
**Changes**: [summary]
### Success Criteria
#### Automated Verification:
- [ ] `make test`
#### Manual Verification:
- [ ] Feature works as expected
**Note**: Pause after automated verification for human manual testing confirmation.
---
## Testing Strategy
## References
If the plan is created inside /Users/fredrick/Desktop/02_Personal/personal-knowledge or is about Fredrick's Personal Knowledge system, append a metadata-only event to:
/Users/fredrick/Desktop/02_Personal/personal-knowledge/queues/events/knowledge-events.jsonl
Event rules:
source_class: "plan" and signal_type: "artifact_completed".source_ref, such as plans/YYYY-MM-DD-description.md.summary short and non-sensitive; do not copy plan body into the event.privacy from the plan context when obvious; otherwise default to sensitive.suggested_route to wiki_candidate or wiki_source_queue.status: "queued".jq -c . /Users/fredrick/Desktop/02_Personal/personal-knowledge/queues/events/knowledge-events.jsonl >/dev/null
For Personal Knowledge repo plans, sync with git:
git status --short
git add plans/YYYY-MM-DD-description.md queues/events/knowledge-events.jsonl
git commit -m "docs(plans): add <description> plan"
git push origin main
For HumanLayer thoughts repos:
humanlayer thoughts sync
Present plan location and ask for feedback. Iterate until satisfied.
Document codebase as-is via parallel sub-agents. Use when researching how code works, mapping architecture, or answering questions about existing systems. Auto-detects thoughts/ directory — if present, integrates historical context and syncs; if absent, skips those steps.
Create the Questions artifact before codebase research, design, planning, or implementation. Use for feature tickets, architecture changes, unclear code work, CRISPY/QRSPI Questions phase, or when the agent needs to define code-answerable unknowns and human questions before researching or planning.
Create a concise Design Concept artifact after questions and codebase research are available. Use for CRISPY/QRSPI Design phase, architecture alignment, feature design discussion, or turning current-state research into a small human-reviewable proposed shape before detailed planning.
Create an interface-first Structure Outline before detailed planning or implementation. Use for CRISPY/QRSPI Structure Outline phase, vertical planning, method signatures, type contracts, API/component boundaries, or converting an accepted design concept into end-to-end slices with verification checkpoints.
Create comprehensive handoff documents for transferring work to another agent session. Use when the user requests to create a handoff, end a session, or document work-in-progress for future continuation.
個人腦力激盪教練。用結構化技巧逼出你自己的想法，而非替你產生點子。 僅在使用者明確說出「腦力激盪」或「brainstorm」時觸發，其他模糊說法（幫我想想、 想不出辦法、有什麼可能）不觸發。 角色：Claude 是 coach，用 probing question 引導，不是 idea generator。