Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
dmlguq456 avatar

Code Plan

  • 1 installs
  • 7 repo stars
  • Updated August 5, 2026
  • dmlguq456/claude_setting

Creates a detailed codebase-grounded implementation plan via a subagent, with QA review levels from quick to adversarial (Korean output).

About

Creates a detailed implementation plan grounded in the actual codebase via a plan-team subagent, with scalable QA review rounds. A developer uses it to plan a coding task before implementation.

  • Delegates plan creation to a plan-team subagent reading real source files
  • Scales QA review from quick to adversarial with an iterative review loop

Code Plan by the numbers

  • 1 all-time installs (skills.sh)
  • Ranked #2,479 of 3,282 Productivity & Planning skills by installs in the Skillselion catalog
  • Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/dmlguq456/claude_setting --skill code-plan

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs1
repo stars7
Last updatedAugust 5, 2026
Repositorydmlguq456/claude_setting

What it does

Creates a detailed codebase-grounded implementation plan via a subagent, with QA review levels from quick to adversarial (Korean output).

Files

SKILL.mdMarkdownGitHub ↗
Caller note: planning benefits from high or xhigh effort; lower effort may miss call sites in cross-file analysis.

Language Rule

  • All user-facing output in natural Korean (no translationese — write Korean natively, don't translate from an English draft).

Pre-Check

Check if a similar plan already exists in .claude_reports/plans/. Behavior depends on plan status:

  • active: Always ask the user — "기존에 진행 중인 plan이 있습니다. 이어서 진행할까요, 새로 만들까요?" Do NOT proceed until confirmed.
  • done/failed: Note it for reference and auto-proceed with new plan creation (no prompt).
  • partial: Auto-create a new plan covering only the failed steps (read failed_steps from plan frontmatter). No prompt.
Record any user-facing pause for the pipeline_summary.md Decision Points table.

Delegate to 기획팀

Invoke the plan-team (기획팀) agent as a subagent with the following prompt:

Plan mode. Create a new implementation plan.

Task: {$ARGUMENTS}
Save English plan to: .claude_reports/plans/{YYYY-MM-DD}_{short-task-name}/plan/plan.md
Date: {YYYY-MM-DD}
{If a done/failed/partial plan exists: "Reference previous plan: [path], status: [status]"}
{If partial: "Failed steps from previous execution: [list from plan frontmatter failed_steps]"}

Read all relevant source files, analyze the current state, and create the plan.
Write the plan files directly. Return ONLY the file paths and a 3-5 line Korean summary. Do NOT return the plan content itself.

The agent writes the plan file directly; the orchestrator only receives paths and a summary.

QA Scaling

If $ARGUMENTS contains --qa quick|light|standard|thorough|adversarial, use that level and strip the flag from the task description. Otherwise, auto-detect from the plan's scope. When qa_level is set in plan frontmatter, it overrides auto-detect.

LevelAuto-detect conditionAction
Quick(manual only — never auto-selected)1× 품질관리팀 (model: "sonnet"), single pass, max 1 review round (no iteration even if 🔴 found — 🔴 are recorded as 미해결 이슈 and loop exits)
Light≤3 steps, mechanical, single-variant1× 품질관리팀 (model: "sonnet")
Standard4-10 steps, logic changes, single module1× 품질관리팀 (default opus)
Thorough>10 steps, cross-module/variant, architectural2-3× 품질관리팀 in parallel: Agent A correctness (opus), B completeness (sonnet), C risk (opus, optional, >15 steps); each writes round_{N}_{focus}.md; all 🔴 issues must be resolved
AdversarialCross-variant (SE+SS+CSS), shared modules (utils/, network.py), or >20 steps with architectural impact — AND Codex availableThorough-level 품질관리팀 + 1× codex-review-team (adversarial-review) in parallel; Codex writes round_{N}_codex.md; all 🔴 from ANY agent (including Codex) must be resolved

Codex availability check: Before selecting Adversarial, run codex --version (suppress stderr). If the command fails or Codex is not authenticated, fall back to Thorough silently. This check is skipped if --qa adversarial is explicitly specified (fail loudly instead).

Post-Plan Review Loop (max 3 revision rounds; quick = 1 round)

The log directory is the task root folder (parent of plan/). Example: .claude_reports/plans/2026-03-18_task/plan/plan.md → log dir is .claude_reports/plans/2026-03-18_task/. Run mkdir -p {log_dir}/_internal/plan_reviews before invoking QA.

Round counting: Initialize round = 0. A round = one plan-team fix → QA review cycle; all parallel Thorough agents count as one round. Increment round only when QA is re-invoked after a revision. "max 3 rounds" means 기획팀 is invoked at most 3 times to fix issues. `quick` mode: max rounds = 1 — after the single review pass, exit regardless of 🔴 (record residuals as 미해결 이슈 and skip the fix-round).

QA level lock: QA level is determined once at loop start; only upward escalation allowed (no downgrade). If --qa was NOT specified, the orchestrator MAY upgrade once (starting round 2) when 🔴 count ≥3 in the one-line verdict (no review file reading needed); round counter does NOT reset. If --qa was manually specified, no change allowed. quick is never auto-upgraded (user opted in for fastest path).

After the 기획팀 agent returns: 1. Assess QA level from plan scope per the QA Scaling table above. 2. Invoke 품질관리팀: Prompt: "Review this plan in plan review mode for feasibility. Plan file: [plan_path]. Write review results to: [log_dir]/_internal/plan_reviews/round_{N}.md. Return ONLY the file path and a one-line verdict."

  • Light: pass model: 'sonnet'. Thorough: 2-3 parallel agents with focus suffix and separate output files; pass model: 'sonnet' for the B (completeness) agent, default opus for A (correctness) and C (risk). Do NOT read the review file unless relaying verdict to user.

3. Check one-line verdict:

  • No 🔴: Loop ends → proceed to Korean Version Generation.
  • 🔴 found AND qa_level == quick: Loop ends (no fix-round). Invoke 기획팀: "Refine mode. Add 🔴 issues from {log_dir}/_internal/plan_reviews/round_1.md to the plan's 리스크 section under ## 미해결 이슈. Return brief Korean summary." Then proceed to Korean Version Generation.
  • 🔴 found: Re-invoke 기획팀: "Refine mode. Plan file: {plan_path}. QA review: {log_dir}/_internal/plan_reviews/round_{N}.md. Fix 🔴 issues. Return only changed steps + brief Korean summary." Increment round, re-invoke 품질관리팀. Repeat until no 🔴 or round >= 3.

4. If 🔴 remain after `round >= 3`: Auto-proceed — invoke 기획팀: "Refine mode. Add remaining 🔴 issues to the plan's 리스크 section under ## 미해결 이슈. Return brief Korean summary." Then report to user: plan path, resolved issues, and unresolved issues with reasons.

Record any user-facing pause (e.g., active-plan ambiguity) so the pipeline skill can surface it in pipeline_summary.md.

Mirror Generation (편집팀 — conditional)

코드 plan 은 _코드 식별자 + 단계 설명_ 묶음 — primary language 는 English (코드 자체가 영문 자연). 한국어 사용자 검토용 mirror 가 보통 필요. 사용자가 영문 plan 만 본다고 명시한 경우 mirror 생성 skip.

Skip condition: 사용자가 영문 plan 만 검토한다고 명시 또는 영문 사용자.

Trigger (default for 한국어 사용자): plan.md 영문 + 한국어 mirror 필요.

모드 A — 영문에서 국문으로 옮기기.
영문 plan 경로: {plan_path}
국문 출력 경로: {same directory}/plan_ko.md
~/.claude/agents/editorial-team.md 의 모드 A 절차를 따른다.
~/.claude/agents/editorial-team.md 의 판교체 회피 절을 강제 적용. 사용자 표기 선호는 `mem profile 02_paper_writing_style` 보조 참조.
코드 식별자·파일 경로·라이브러리 이름은 영어 그대로, 그 외 일반 표현은 한국어로.
section 제목 매핑: Goals → 목표, Current State → 현황 분석, Change Plan → 변경 계획, Risks → 리스크, Verification → 검증 방법.
완료 시 파일 경로 + 한국어 요약 3-5 줄 + 의도적으로 한 표기 결정 한두 개만 돌려준다.

Then report to the user: plan path(s) + summary + QA verdict.

Task

$ARGUMENTS

Related skills

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.