improve(v3.2.90): START_PROMPT.md を ClaudeOS v8.5 Ultimate に刷新 - #252
Merged
Merged
Conversation
- Claude/templates/claude/START_PROMPT.md を全面刷新 - 6ヶ月フェーズ制御 (Build/Quality/Stabilize/Release) + KPIスコア連動ループ - フェーズ別時間配分黄金比 + GitHub Actions 完全連動テンプレート - AI Dev Factory Issue テンプレート収録 - New-CronSchedule.ps1 の Invoke-SyncStartPrompt で全プロジェクトへ配備可 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
📝 WalkthroughWalkthroughClaudeOS v8.5 Ultimate（v3.2.90）向けドキュメントとテンプレートの大規模刷新。START_PROMPT を全面書き換え（実行制約、6ヶ月フェーズ制御、KPIスコア判定、簡易 state.json、簡略化された CI/Projects/AI Dev Factory テンプレ）。README/CHANGELOG 更新、Cron同期とログ処理・設定検証の安全性向上。 Changes
Sequence Diagram(s)sequenceDiagram
participant Monitor as Monitor
participant ClaudeOS as ClaudeOS
participant GitHubActions as "GitHub Actions"
participant GitHubProjects as "GitHub Projects"
participant AI_DevFactory as "AI Dev Factory"
Monitor->>ClaudeOS: 監視データ送信 / KPI 評価要求
ClaudeOS->>ClaudeOS: KPI 判定・フェーズ/ループ決定
ClaudeOS->>GitHubActions: PR/CI トリガー（lint/test/build + KPI 更新）
GitHubActions-->>ClaudeOS: CI 結果 / KPI 更新結果
ClaudeOS->>GitHubProjects: フェーズ→ボードのステータス同期
ClaudeOS->>AI_DevFactory: CI失敗/KPI未達/セキュリティで Issue 作成
AI_DevFactory-->>GitHubActions: Issue に基づく修復ワークフロー提案
ClaudeOS->>Monitor: 状態報告 / 次アクション決定
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Comment |
- Invoke-SyncMenu: Select-Project (1件選択) → 全プロジェクト列挙 + 個別 Y/N 確認に変更 - Get-ProjectList で Linux から全プロジェクトを取得し foreach で順次処理 - 同期済み件数 / スキップ件数をサマリー表示 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@Claude/templates/claude/START_PROMPT.md`:
- Around line 114-152: The KPI section in state.json (the "kpi" object) lacks
explicit target thresholds so goals can't be evaluated from the file alone; add
target fields (e.g., success_rate_target, ci_success_rate_target,
test_pass_rate_target, review_blocker_target, security_issue_target) at the same
level as existing counters and mirror the placement of
execution.max_duration_minutes so consumers can read both limits and targets
from state.json; update any code or schema references that read "kpi" to expect
these new target properties (look for uses of "kpi",
"execution.max_duration_minutes", and functions/classes that load state.json).
- Around line 156-195: The "GitHub Actions（完全連動）" snippet in START_PROMPT.md is
misleading vs the real CI; either update the YAML sample to match the actual
workflow (replace ubuntu-latest/Node-centric jobs like build-test and
auto-repair with the real jobs used in your repo: Windows runners, Pester tests,
schema validation, README consistency checks, and the true job names/steps), or
change the heading and paragraph to clearly label it as a generic "参考テンプレート" and
add a short note that it is not synchronized with .github/workflows/ci.yml;
locate the section containing the build-test and auto-repair job blocks and
adjust the text/YAML accordingly so the template reflects real CI behavior or is
clearly marked as an example.
In `@README.md`:
- Around line 10-11: The README header was updated to advertise "ClaudeOS v8.5
Ultimate" but the body still contains legacy descriptions; update the README
sections that describe Loop Definitions, Operation Specifications, and KPI
Definitions to match v8.5 Ultimate constraints and flows (including the 6-month
phase control, routine optimizations, KPI-linked feedback loop, and full GitHub
Actions/AI Dev Factory integration), ensure setup and architecture steps reflect
the new self-autonomous OS template and any changed quality gates, and add brief
notes or links to START_PROMPT.md and CHANGELOG.md where deeper technical
details live so the README remains the single-source-of-truth for external
users.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 4e9728f1-9b6e-4dfd-bd9e-b954b7ff7739
📒 Files selected for processing (3)
CHANGELOG.mdClaude/templates/claude/START_PROMPT.mdREADME.md
📜 Review details
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)
- GitHub Check: Agent
- GitHub Check: PSScriptAnalyzer
- GitHub Check: test-and-validate
🧰 Additional context used
📓 Path-based instructions (2)
**/README.md
📄 CodeRabbit inference engine (CLAUDE.md)
**/README.md: Record that CI is not yet set up if quality gates (lint, unit test, build, dependency/security scan) are unavailable
Update README.md whenever utilizer-facing features, setup procedures, architecture, or quality gates change
Treat README as external-facing truth; do not publish or use a README that cannot withstand external scrutiny
Files:
README.md
**/*.{js,ts,jsx,tsx,py,java,go,rs,cs,cpp,c,rb,php,md}
📄 CodeRabbit inference engine (CLAUDE.md)
Create GitHub Issues for KPI unmet, CI failures, review findings, TODO/FIXME detection, insufficient tests, or security concerns
Files:
README.mdCHANGELOG.mdClaude/templates/claude/START_PROMPT.md
🧠 Learnings (4)
📓 Common learnings
Learnt from: CR
Repo: Kensan196948G/ClaudeCode-StartUpTools-New PR: 0
File: Claude/templates/claude/CLAUDE.md:0-0
Timestamp: 2026-04-23T11:14:27.821Z
Learning: Applies to Claude/templates/claude/**/README.md : Update README.md whenever: user-facing functionality changes, setup procedures change, architecture changes, or quality gates change
📚 Learning: 2026-04-23T11:14:27.821Z
Learnt from: CR
Repo: Kensan196948G/ClaudeCode-StartUpTools-New PR: 0
File: Claude/templates/claude/CLAUDE.md:0-0
Timestamp: 2026-04-23T11:14:27.821Z
Learning: Applies to Claude/templates/claude/**/README.md : Update README.md whenever: user-facing functionality changes, setup procedures change, architecture changes, or quality gates change
Applied to files:
README.mdCHANGELOG.mdClaude/templates/claude/START_PROMPT.md
📚 Learning: 2026-04-22T08:01:46.016Z
Learnt from: CR
Repo: Kensan196948G/ClaudeCode-StartUpTools-New PR: 0
File: CLAUDE.md:0-0
Timestamp: 2026-04-22T08:01:46.016Z
Learning: Applies to **/README.md : Update README.md whenever utilizer-facing features, setup procedures, architecture, or quality gates change
Applied to files:
README.md
📚 Learning: 2026-04-23T11:14:27.821Z
Learnt from: CR
Repo: Kensan196948G/ClaudeCode-StartUpTools-New PR: 0
File: Claude/templates/claude/CLAUDE.md:0-0
Timestamp: 2026-04-23T11:14:27.821Z
Learning: Applies to Claude/templates/claude/**/state.json : Use state.json as the single source of truth for project goals and KPI targets with defined success_rate_target and execution max_duration_minutes
Applied to files:
Claude/templates/claude/START_PROMPT.md
🔇 Additional comments (1)
CHANGELOG.md (1)
5-25: 変更履歴の追記は明確で問題ありません。v3.2.90 の要点が簡潔にまとまっており、リリースノートとして読みやすいです。
Comment on lines
+114
to
152
| # 🧠 ■ state.json（完全版） | ||
|
|
||
| ```json | ||
| { | ||
| "project": { | ||
| "name": "sample-project", | ||
| "mode": "autonomous" | ||
| "name": "project-name", | ||
| "start_date": "2026-01-01", | ||
| "release_deadline": "2026-07-01" | ||
| }, | ||
| "goal": { | ||
| "title": "自律開発最適化", | ||
| "description": "品質と安定性を維持しながら継続的に改善する" | ||
| "phase": { | ||
| "current": "build", | ||
| "week": 1 | ||
| }, | ||
| "kpi": { | ||
| "success_rate_target": 0.9, | ||
| "test_pass_rate_target": 0.95, | ||
| "review_blocker_target": 0, | ||
| "security_blocker_target": 0, | ||
| "ci_stability_target": 0.95 | ||
| "ci_success_rate": 0.0, | ||
| "test_pass_rate": 0.0, | ||
| "review_blocker_count": 0, | ||
| "security_issue_count": 0 | ||
| }, | ||
| "execution": { | ||
| "max_duration_minutes": 300, | ||
| "cooldown_minutes_min": 5, | ||
| "cooldown_minutes_max": 15, | ||
| "retry_limit_ci": 15, | ||
| "same_root_cause_limit": 3 | ||
| "loop_count": 0, | ||
| "max_loops": 3, | ||
| "ci_retry_limit": 5, | ||
| "same_error_limit": 2 | ||
| }, | ||
| "automation": { | ||
| "auto_issue_generation": true, | ||
| "auto_project_sync": true, | ||
| "self_evolution": true, | ||
| "auto_priority_scoring": true, | ||
| "auto_repair": true | ||
| "status": { | ||
| "current_phase": "monitor", | ||
| "stable": false | ||
| }, | ||
| "priority": { | ||
| "weights": { | ||
| "security": 100, | ||
| "ci_failure": 90, | ||
| "data_risk": 85, | ||
| "test_failure": 75, | ||
| "review_findings": 70, | ||
| "kpi_gap": 65, | ||
| "technical_debt": 40, | ||
| "minor_ux": 20 | ||
| }, | ||
| "current_top_reason": "ci_failure" | ||
| "score": 0 | ||
| }, | ||
| "learning": { | ||
| "failure_patterns": [], | ||
| "success_patterns": [], | ||
| "blocked_patterns": [], | ||
| "preferred_fix_order": ["security", "ci", "test", "review", "refactor"] | ||
| }, | ||
| "github": { | ||
| "default_branch": "main", | ||
| "require_pr": true, | ||
| "require_codex_review": true, | ||
| "require_actions_success": true, | ||
| "project_sync_enabled": true | ||
| }, | ||
| "status": { | ||
| "stable": false, | ||
| "blocked": false, | ||
| "current_phase": "monitor", | ||
| "last_updated": "YYYY-MM-DDTHH:MM:SSZ" | ||
| "success_patterns": [] | ||
| } | ||
| } | ||
| ``` |
There was a problem hiding this comment.
state.json のKPI目標定義が不足しています。
Line 127-146 の構造はカウンタ中心ですが、成功率の目標値（success_rate_target）がないため、KPIに基づく判定基準を state.json 単体で完結できません。execution.max_duration_minutes はあるので、同レベルで目標値も明示してください。
Based on learnings, "Use state.json as the single source of truth for project goals and KPI targets with defined success_rate_target and execution max_duration_minutes".
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@Claude/templates/claude/START_PROMPT.md` around lines 114 - 152, The KPI
section in state.json (the "kpi" object) lacks explicit target thresholds so
goals can't be evaluated from the file alone; add target fields (e.g.,
success_rate_target, ci_success_rate_target, test_pass_rate_target,
review_blocker_target, security_issue_target) at the same level as existing
counters and mirror the placement of execution.max_duration_minutes so consumers
can read both limits and targets from state.json; update any code or schema
references that read "kpi" to expect these new target properties (look for uses
of "kpi", "execution.max_duration_minutes", and functions/classes that load
state.json).
Comment on lines
+156
to
195
| # ⚙️ ■ GitHub Actions（完全連動） | ||
|
|
||
| ```yaml | ||
| name: ClaudeOS CI Manager | ||
|
|
||
| on: | ||
| push: | ||
| branches: [main, develop, 'feature/**'] | ||
| branches: [main, develop, "feature/**"] | ||
| pull_request: | ||
| branches: [main, develop] | ||
| branches: [main] | ||
| workflow_dispatch: | ||
|
|
||
| jobs: | ||
| build-test-review: | ||
| build-test: | ||
| runs-on: ubuntu-latest | ||
| steps: | ||
| - name: Checkout | ||
| uses: actions/checkout@v4 | ||
| - uses: actions/checkout@v4 | ||
|
|
||
| - name: Setup Node | ||
| uses: actions/setup-node@v4 | ||
| with: | ||
| node-version: '20' | ||
| node-version: "20" | ||
|
|
||
| - name: Install | ||
| run: | | ||
| if [ -f package-lock.json ]; then npm ci; else npm install; fi | ||
| - run: npm ci || npm install | ||
| - run: npm run lint --if-present | ||
| - run: npm test --if-present | ||
| - run: npm run build --if-present | ||
|
|
||
| - name: Lint | ||
| run: npm run lint --if-present | ||
|
|
||
| - name: Test | ||
| run: npm test --if-present | ||
|
|
||
| - name: Build | ||
| run: npm run build --if-present | ||
|
|
||
| - name: Security Audit | ||
| run: npm audit --audit-level=high || true | ||
|
|
||
| - name: Archive logs | ||
| if: always() | ||
| uses: actions/upload-artifact@v4 | ||
| with: | ||
| name: ci-logs | ||
| path: . | ||
| - name: KPI Update | ||
| run: | | ||
| echo "CI SUCCESS" > ci_result.txt | ||
|
|
||
| issue-factory: | ||
| auto-repair: | ||
| if: failure() | ||
| needs: build-test-review | ||
| runs-on: ubuntu-latest | ||
| steps: | ||
| - name: Checkout | ||
| uses: actions/checkout@v4 | ||
|
|
||
| - name: Create Failure Report | ||
| - name: Create Issue | ||
| run: | | ||
| mkdir -p .claudeos | ||
| echo "CI failed - generate issue" > .claudeos/ci_failure_report.txt | ||
|
|
||
| - name: Upload Failure Report | ||
| uses: actions/upload-artifact@v4 | ||
| with: | ||
| name: ci-failure-report | ||
| path: .claudeos/ci_failure_report.txt | ||
| echo "CI failure detected" > issue.txt | ||
| ``` |
There was a problem hiding this comment.
「GitHub Actions 完全連動」の記述が実ワークフローと乖離しています。
このテンプレートは ubuntu-latest + Node中心 + 疑似的な echo ステップですが、実際の .github/workflows/ci.yml は Windows/Pester/スキーマ検証/README整合チェック中心です。完全連動 と表現するなら、現行CIに合わせた実運用準拠のサンプルへ更新するか、参考テンプレートである旨を明記してください。
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@Claude/templates/claude/START_PROMPT.md` around lines 156 - 195, The "GitHub
Actions（完全連動）" snippet in START_PROMPT.md is misleading vs the real CI; either
update the YAML sample to match the actual workflow (replace
ubuntu-latest/Node-centric jobs like build-test and auto-repair with the real
jobs used in your repo: Windows runners, Pester tests, schema validation, README
consistency checks, and the true job names/steps), or change the heading and
paragraph to clearly label it as a generic "参考テンプレート" and add a short note that
it is not synchronized with .github/workflows/ci.yml; locate the section
containing the build-test and auto-repair job blocks and adjust the text/YAML
accordingly so the template reflects real CI behavior or is clearly marked as an
example.
Comment on lines
+10
to
+11
| > **🚀 v3.2.90 — START_PROMPT.md を ClaudeOS v8.5 Ultimate に刷新** | ||
| > Routines最適化・6ヶ月フェーズ制御・KPI連動ループ・GitHub Actions完全連動・AI Dev Factory を統合した完全自律開発 OS テンプレートに更新。詳細は [`CHANGELOG.md`](./CHANGELOG.md) を参照。 |
There was a problem hiding this comment.
v8.5告知に対して README 本文仕様の同期が不足しています。
Line 10-11 と Line 30 で v8.5 Ultimate へ更新した旨を示していますが、README 内の詳細説明（ループ定義・運用仕様）が旧記述のまま残っており、読者に誤解を与えます。v8.5 の実際の制約・フロー・KPI定義へ本文側も合わせて更新してください。
As per coding guidelines, "Update README.md whenever utilizer-facing features, setup procedures, architecture, or quality gates change" and "Treat README as external-facing truth; do not publish or use a README that cannot withstand external scrutiny".
Also applies to: 30-30
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@README.md` around lines 10 - 11, The README header was updated to advertise
"ClaudeOS v8.5 Ultimate" but the body still contains legacy descriptions; update
the README sections that describe Loop Definitions, Operation Specifications,
and KPI Definitions to match v8.5 Ultimate constraints and flows (including the
6-month phase control, routine optimizations, KPI-linked feedback loop, and full
GitHub Actions/AI Dev Factory integration), ensure setup and architecture steps
reflect the new self-autonomous OS template and any changed quality gates, and
add brief notes or links to START_PROMPT.md and CHANGELOG.md where deeper
technical details live so the README remains the single-source-of-truth for
external users.
There was a problem hiding this comment.
Pull request overview
Claude/templates/claude/START_PROMPT.md を ClaudeOS v8.5 Ultimate テンプレートとして刷新し、ドキュメント側のバージョン表記（README/CHANGELOG）を v3.2.90 に更新するPRです。
Changes:
Claude/templates/claude/START_PROMPT.mdの内容を v8.5 Ultimate 仕様へ全面差し替えCHANGELOG.mdに v3.2.90 エントリを追加README.mdのバージョン表記と冒頭サマリを v3.2.90 に更新
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 6 comments.
| File | Description |
|---|---|
| README.md | v3.2.90 の告知・バージョン表更新 |
| Claude/templates/claude/START_PROMPT.md | v8.5 Ultimate の開始プロンプト（KPI/フェーズ/Actions/Factory等）へ刷新 |
| CHANGELOG.md | v3.2.90 のリリースノート追加 |
Comment on lines
+1
to
+5
| # 🚀 ClaudeOS v8.5 Ultimate（完全統合版） | ||
|
|
||
| ## セッション開始処理 | ||
| ## 🎯 CTO完全自立型開発OS | ||
|
|
||
| 自律実行は Linux cron（月〜土、プロジェクト別スケジュール・300分制限）が担う。 | ||
| セッション開始時に loop や schedule コマンドの登録は不要。 | ||
|
|
||
| Codex が利用可能な場合のみ実行する。**Codex が使えなくても自律開発は止めない。** | ||
|
|
||
| ``` | ||
| /codex:setup | ||
| /codex:status | ||
| ``` | ||
|
|
||
| - Codex 認証済み → review 強化レイヤーとして活用 | ||
| - Codex 未認証 / 使えない → このステップをスキップして自律開発を開始する | ||
| - **Codex の有無は停止条件ではない** | ||
|
|
||
| ## PROMPT_BODY | ||
|
|
||
| # ClaudeOS v8 完全無人運用版 | ||
|
|
||
| ## Autonomous Operations Edition + AI Dev Factory + Priority Intelligence + Full CI Automation + Auto Loop Intelligence + CodeRabbit Review Integration | ||
|
|
||
| ## 指示確認プロトコル（必須） | ||
|
|
||
| 作業を開始する前に、以下の確認チェックリストを出力してください。 | ||
|
|
||
| ``` | ||
| === ClaudeOS v8 指示確認 === | ||
| [✅/❌] 01 セッション開始・実行方針 | ||
| [✅/❌] 02 コア構造・マルチプロジェクト・Goal Driven | ||
| [✅/❌] 03 state.json スキーマ・優先順位AI | ||
| [✅/❌] 04 Agent Teams 定義・起動順序 | ||
| [✅/❌] 05 Codex統合（任意）・Debug原則 | ||
| [✅/❌] 06 CI Manager・GitHub Actions | ||
| [✅/❌] 07 AI Dev Factory・GitHub Projects連携 | ||
| [✅/❌] 08 ループ制御・WorkTree・Token・時間・STABLE・禁止事項 | ||
| [✅/❌] 09 終了処理・最終報告・可視化・ドキュメント運用 | ||
| ============================== | ||
| ``` | ||
| 以降は全てCTO全権委任で自律開発を開始してください。 | ||
| 全てをCTOの判断に任せます。CTO全権委任で自律開発を続けてください。 | ||
| Auto Mode による自律開発を実行してください。 | ||
| 全項目 ✅ を確認してから Monitor フェーズを開始すること。 | ||
|
|
||
| **設定された時間内で Monitor → Development → Verify → Improvement をループ（ループ回数は CTO 判断で OK）で進めてください。** | ||
| ループ回数は KPI 状態・CI 状態・進捗により動的に決定する。 | ||
| **Routines最適化 + state.json + GitHub Actions + Projects 完全連動** |
There was a problem hiding this comment.
この START_PROMPT.md は Claude/templates/claude/Build-StartPrompt.ps1 により instructions/ から自動生成され、起動時にも Start-ClaudeCode.ps1 が再ビルドを実行します。現状、instructions 側が更新されていないため、ユーザーが起動すると本PRの内容がすぐ上書きされて反映されません。v8.5 Ultimate を配布したい場合は、START_PROMPT.md ではなく instructions/ のソース一式を更新するか、再ビルド処理を無効化/条件化してください。
Comment on lines
+114
to
151
| # 🧠 ■ state.json（完全版） | ||
|
|
||
| ```json | ||
| { | ||
| "project": { | ||
| "name": "sample-project", | ||
| "mode": "autonomous" | ||
| "name": "project-name", | ||
| "start_date": "2026-01-01", | ||
| "release_deadline": "2026-07-01" | ||
| }, | ||
| "goal": { | ||
| "title": "自律開発最適化", | ||
| "description": "品質と安定性を維持しながら継続的に改善する" | ||
| "phase": { | ||
| "current": "build", | ||
| "week": 1 | ||
| }, | ||
| "kpi": { | ||
| "success_rate_target": 0.9, | ||
| "test_pass_rate_target": 0.95, | ||
| "review_blocker_target": 0, | ||
| "security_blocker_target": 0, | ||
| "ci_stability_target": 0.95 | ||
| "ci_success_rate": 0.0, | ||
| "test_pass_rate": 0.0, | ||
| "review_blocker_count": 0, | ||
| "security_issue_count": 0 | ||
| }, | ||
| "execution": { | ||
| "max_duration_minutes": 300, | ||
| "cooldown_minutes_min": 5, | ||
| "cooldown_minutes_max": 15, | ||
| "retry_limit_ci": 15, | ||
| "same_root_cause_limit": 3 | ||
| "loop_count": 0, | ||
| "max_loops": 3, | ||
| "ci_retry_limit": 5, | ||
| "same_error_limit": 2 | ||
| }, | ||
| "automation": { | ||
| "auto_issue_generation": true, | ||
| "auto_project_sync": true, | ||
| "self_evolution": true, | ||
| "auto_priority_scoring": true, | ||
| "auto_repair": true | ||
| "status": { | ||
| "current_phase": "monitor", | ||
| "stable": false | ||
| }, | ||
| "priority": { | ||
| "weights": { | ||
| "security": 100, | ||
| "ci_failure": 90, | ||
| "data_risk": 85, | ||
| "test_failure": 75, | ||
| "review_findings": 70, | ||
| "kpi_gap": 65, | ||
| "technical_debt": 40, | ||
| "minor_ux": 20 | ||
| }, | ||
| "current_top_reason": "ci_failure" | ||
| "score": 0 | ||
| }, | ||
| "learning": { | ||
| "failure_patterns": [], | ||
| "success_patterns": [], | ||
| "blocked_patterns": [], | ||
| "preferred_fix_order": ["security", "ci", "test", "review", "refactor"] | ||
| }, | ||
| "github": { | ||
| "default_branch": "main", | ||
| "require_pr": true, | ||
| "require_codex_review": true, | ||
| "require_actions_success": true, | ||
| "project_sync_enabled": true | ||
| }, | ||
| "status": { | ||
| "stable": false, | ||
| "blocked": false, | ||
| "current_phase": "monitor", | ||
| "last_updated": "YYYY-MM-DDTHH:MM:SSZ" | ||
| "success_patterns": [] | ||
| } | ||
| } |
There was a problem hiding this comment.
ここで提示している state.json（完全版） の例が、リポジトリの state.schema.json（必須: goal/kpi/execution/stable/token/effort_strategy/compact/notification 等）とフィールド構成が一致していません（例: project/phase/status/priority.score などはスキーマに無い）。このままだとテンプレを信じて state.json を生成した際に hooks/検証（CI の schema validation）と齟齬が出るため、スキーマに沿った例に更新するか「擬似例でありスキーマ非準拠」と明記してください。
|
|
||
| issue-factory: | ||
| auto-repair: | ||
| if: failure() |
There was a problem hiding this comment.
GitHub Actions テンプレ内の auto-repair job が needs を持たないまま if: failure() を付けています。この形だと build job の結果を待たずに評価されて skip されやすく、意図した「失敗時のみ実行」になりません。auto-repair に needs: build-test を付けた上で if: failure()（または if: ${{ needs.build-test.result == 'failure' }}）にする構成へ修正してください。
Suggested change
| if: failure() | |
| needs: build-test | |
| if: ${{ needs.build-test.result == 'failure' }} |
Comment on lines
+120
to
+121
| "start_date": "2026-01-01", | ||
| "release_deadline": "2026-07-01" |
There was a problem hiding this comment.
例の start_date / release_deadline が 2026 の固定日付になっており、テンプレとしてはすぐ陳腐化します。YYYY-MM-DD のようなプレースホルダにするか、「例なので任意の日付へ置換する」旨を明記した方が誤用を防げます。
Suggested change
| "start_date": "2026-01-01", | |
| "release_deadline": "2026-07-01" | |
| "start_date": "YYYY-MM-DD", | |
| "release_deadline": "YYYY-MM-DD" |
| > **🎨 v3.2.89 — メニュー大幅刷新 + ログ監視タブ発火タイミング修正** | ||
| > `Start-Menu.ps1` を PowerShell 7 カラー + Unicode ボーダー + 絵文字アイコンで全面刷新。`Watch-ClaudeLog.ps1` の `stdbuf` 依存を除去し cron 発火タイミング検知遅延を解消。詳細は [`CHANGELOG.md`](./CHANGELOG.md) を参照。 | ||
| > **🚀 v3.2.90 — START_PROMPT.md を ClaudeOS v8.5 Ultimate に刷新** | ||
| > Routines最適化・6ヶ月フェーズ制御・KPI連動ループ・GitHub Actions完全連動・AI Dev Factory を統合した完全自律開発 OS テンプレートに更新。詳細は [`CHANGELOG.md`](./CHANGELOG.md) を参照。 |
There was a problem hiding this comment.
README の新しいサマリ行で「Routines最適化」となっていますが、CHANGELOG 側は「Routines 最適化」とスペース入りで表記されており表記ゆれがあります。どちらかに揃えると検索性/一貫性が上がります。
Suggested change
| > Routines最適化・6ヶ月フェーズ制御・KPI連動ループ・GitHub Actions完全連動・AI Dev Factory を統合した完全自律開発 OS テンプレートに更新。詳細は [`CHANGELOG.md`](./CHANGELOG.md) を参照。 | |
| > Routines 最適化・6ヶ月フェーズ制御・KPI連動ループ・GitHub Actions完全連動・AI Dev Factory を統合した完全自律開発 OS テンプレートに更新。詳細は [`CHANGELOG.md`](./CHANGELOG.md) を参照。 |
Comment on lines
8
to
+11
| > v3.1.0 より、Codex CLI / GitHub Copilot CLI の起動メニュー (S2/S3/L2/L3) は削除されました。本ツールは **Claude Code 専用の自律開発ランチャー** として位置づけを明確化し、Linux crontab 連携・セッション情報タブ・Statusline グローバル適用などの新機能に投資が集中しています。 | ||
|
|
||
| > **🎨 v3.2.89 — メニュー大幅刷新 + ログ監視タブ発火タイミング修正** | ||
| > `Start-Menu.ps1` を PowerShell 7 カラー + Unicode ボーダー + 絵文字アイコンで全面刷新。`Watch-ClaudeLog.ps1` の `stdbuf` 依存を除去し cron 発火タイミング検知遅延を解消。詳細は [`CHANGELOG.md`](./CHANGELOG.md) を参照。 | ||
| > **🚀 v3.2.90 — START_PROMPT.md を ClaudeOS v8.5 Ultimate に刷新** | ||
| > Routines最適化・6ヶ月フェーズ制御・KPI連動ループ・GitHub Actions完全連動・AI Dev Factory を統合した完全自律開発 OS テンプレートに更新。詳細は [`CHANGELOG.md`](./CHANGELOG.md) を参照。 |
There was a problem hiding this comment.
今回 v8.5 Ultimate への刷新を謳っていますが、README 冒頭側の説明ではまだ ClaudeOS v8 として記載されています。バージョン表記が混在するので、v8.5 を正式に名乗るなら冒頭の v8 記載も合わせて更新してください。
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@scripts/main/New-CronSchedule.ps1`:
- Around line 241-257: The summary counter is incremented unconditionally so a
failed sync still increases $synced; update the loop that iterates $projects and
calls Invoke-SyncStartPrompt so it only increments $synced when
Invoke-SyncStartPrompt reports success (e.g., returns true or no error), add a
new $failed (or $error) counter to track failures, increment $skipped only on
explicit skips, and adjust the final Write-Host summary to display synced,
failed, and skipped counts; use the existing variables $synced, $skipped, and
the Invoke-SyncStartPrompt call to locate the change.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 1c8f1cfa-9249-4315-a5e9-477f8f8b1b17
📒 Files selected for processing (1)
scripts/main/New-CronSchedule.ps1
📜 Review details
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: test-and-validate
🧰 Additional context used
🧠 Learnings (1)
📓 Common learnings
Learnt from: CR
Repo: Kensan196948G/ClaudeCode-StartUpTools-New PR: 0
File: Claude/templates/claude/CLAUDE.md:0-0
Timestamp: 2026-04-23T11:14:27.821Z
Learning: Applies to Claude/templates/claude/**/README.md : Update README.md whenever: user-facing functionality changes, setup procedures change, architecture changes, or quality gates change
Learnt from: CR
Repo: Kensan196948G/ClaudeCode-StartUpTools-New PR: 0
File: CLAUDE.md:0-0
Timestamp: 2026-04-22T08:01:46.016Z
Learning: Organize `.claude/claudeos` directory as the standard location for agents, skills, commands, rules, hooks, scripts, contexts, examples, mcp-configs, and kernel documents
Comment on lines
+241
to
257
| $synced = 0 | ||
| $skipped = 0 | ||
|
|
||
| foreach ($p in $projects) { | ||
| $confirm = Read-Host " [$p] 同期しますか? [y/N]" | ||
| if ($confirm -match '^[yY]') { | ||
| Invoke-SyncStartPrompt -Project $p | ||
| $synced++ | ||
| } else { | ||
| Write-Host " [$p] スキップ" -ForegroundColor DarkGray | ||
| $skipped++ | ||
| } | ||
| } | ||
|
|
||
| Write-Host "" | ||
| Write-Host " 完了: 同期 $synced 件 / スキップ $skipped 件" -ForegroundColor Cyan | ||
| } |
There was a problem hiding this comment.
同期件数サマリーが実結果とずれる可能性があります。
Line 247 の同期が失敗しても Line 248 で synced を加算しているため、完了表示が過大になります。成否でカウントを分けてください。
修正案（成否を分離して集計）
- $synced = 0
- $skipped = 0
+ $synced = 0
+ $skipped = 0
+ $failed = 0
foreach ($p in $projects) {
$confirm = Read-Host " [$p] 同期しますか? [y/N]"
if ($confirm -match '^[yY]') {
Invoke-SyncStartPrompt -Project $p
- $synced++
+ if ($LASTEXITCODE -eq 0) {
+ $synced++
+ } else {
+ Write-Host " [$p] 同期失敗" -ForegroundColor Red
+ $failed++
+ }
} else {
Write-Host " [$p] スキップ" -ForegroundColor DarkGray
$skipped++
}
}
Write-Host ""
- Write-Host " 完了: 同期 $synced 件 / スキップ $skipped 件" -ForegroundColor Cyan
+ Write-Host " 完了: 同期 $synced 件 / 失敗 $failed 件 / スキップ $skipped 件" -ForegroundColor Cyan🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@scripts/main/New-CronSchedule.ps1` around lines 241 - 257, The summary
counter is incremented unconditionally so a failed sync still increases $synced;
update the loop that iterates $projects and calls Invoke-SyncStartPrompt so it
only increments $synced when Invoke-SyncStartPrompt reports success (e.g.,
returns true or no error), add a new $failed (or $error) counter to track
failures, increment $skipped only on explicit skips, and adjust the final
Write-Host summary to display synced, failed, and skipped counts; use the
existing variables $synced, $skipped, and the Invoke-SyncStartPrompt call to
locate the change.
- ConfigLoader.ps1: null 中間キーで $obj=$null+break 早期リターン — 誤オブジェクトマスクを防止 - ConfigSchema.ps1: env 型チェックを負の排除 → 正の型チェック (pscustomobject/IDictionary) に変更 - cron-launcher.sh: pipe-pane に sed ANSIフィルタ追加 + || true 廃止 → WARN ログ出力に変更 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
There was a problem hiding this comment.
🧹 Nitpick comments (1)
Claude/templates/linux/cron-launcher.sh (1)
258-263: ANSI除去ルールの重複定義は将来の乖離リスクがあります。Line 258-263 の
sedフィルタと、scripts/tools/Watch-ClaudeLog.ps1（Line 205-215）のフィルタが重複しています。どちらか一方を正としてもう一方は最小化する方針にすると、将来のメンテ時に挙動差分を防げます。🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@Claude/templates/linux/cron-launcher.sh` around lines 258 - 263, The ANSI-stripping sed expression used inline in the tmux pipe-pane command (the line invoking tmux pipe-pane -t "$TMUX_SESSION" -o "sed 's/.*\r//; s/\x1b\][^\x07]*\x07//g; s/\x1b\[[0-9;?]*[a-zA-Z]//g; s/\x1b.//g' >> '$LOG_FILE'") is duplicated with the filter in scripts/tools/Watch-ClaudeLog.ps1; extract the canonical regex into a single reusable place (e.g., a small script/utility like scripts/strip_ansi.sh or a well-documented constant) and replace the inline sed call in cron-launcher.sh and the PowerShell filter in Watch-ClaudeLog.ps1 to invoke that shared utility so both consumers use the exact same rule; keep the command semantics (read from stdin, write sanitized output) and update any invocation to preserve redirection into $LOG_FILE.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@Claude/templates/linux/cron-launcher.sh`:
- Around line 258-263: The ANSI-stripping sed expression used inline in the tmux
pipe-pane command (the line invoking tmux pipe-pane -t "$TMUX_SESSION" -o "sed
's/.*\r//; s/\x1b\][^\x07]*\x07//g; s/\x1b\[[0-9;?]*[a-zA-Z]//g; s/\x1b.//g' >>
'$LOG_FILE'") is duplicated with the filter in
scripts/tools/Watch-ClaudeLog.ps1; extract the canonical regex into a single
reusable place (e.g., a small script/utility like scripts/strip_ansi.sh or a
well-documented constant) and replace the inline sed call in cron-launcher.sh
and the PowerShell filter in Watch-ClaudeLog.ps1 to invoke that shared utility
so both consumers use the exact same rule; keep the command semantics (read from
stdin, write sanitized output) and update any invocation to preserve redirection
into $LOG_FILE.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 15a16c65-b7a4-42f6-92e0-15b9d09b171f
📒 Files selected for processing (3)
Claude/templates/linux/cron-launcher.shscripts/lib/ConfigLoader.ps1scripts/lib/ConfigSchema.ps1
📜 Review details
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: test-and-validate
🧰 Additional context used
📓 Path-based instructions (1)
{**/.github/workflows/*.{yml,yaml},**/ci.{yml,yaml},**/*.sh,**/.gitlab-ci.yml,.circleci/**/*.yml}
📄 CodeRabbit inference engine (AGENTS.md)
CI failures must always be treated as failures; success faking (e.g.,
|| true) is prohibited
Files:
Claude/templates/linux/cron-launcher.sh
🧠 Learnings (1)
📓 Common learnings
Learnt from: CR
Repo: Kensan196948G/ClaudeCode-StartUpTools-New PR: 0
File: Claude/templates/claude/CLAUDE.md:0-0
Timestamp: 2026-04-23T11:14:27.821Z
Learning: Applies to Claude/templates/claude/**/README.md : Update README.md whenever: user-facing functionality changes, setup procedures change, architecture changes, or quality gates change
🔇 Additional comments (3)
scripts/lib/ConfigSchema.ps1 (1)
134-135:envの型制約強化は適切です。
tools.<toolName>.envをオブジェクト系に限定したことで、null/文字列/配列の混入を早期検知でき、設定破損の防止に有効です。scripts/lib/ConfigLoader.ps1 (1)
147-149: 中間キー欠落時の null 安全化は良い修正です。キー経路が途中で切れているケースで明示的に打ち切り、最終参照もガードしているため、NPE相当の失敗を防げています。
Also applies to: 153-154
Claude/templates/linux/cron-launcher.sh (1)
263-265:tmux pipe-pane失敗時の扱い改善は良いです。
|| trueをやめて、警告を stderr とログへ明示出力する形になっており、障害の可観測性が上がっています。As per coding guidelines:
CI failures must always be treated as failures; success faking (e.g., || true) is prohibited.
This was referenced Apr 27, 2026
This was referenced May 11, 2026
11 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
概要
Claude/templates/claude/START_PROMPT.mdを ClaudeOS v8.5 Ultimate（完全統合版） に全面刷新。変更内容
Claude/templates/claude/START_PROMPT.mdCHANGELOG.mdREADME.mdv8.5 Ultimate の主な新規要素
配備方法
New-CronSchedule.ps1メニュー →[7] START_PROMPT を同期を実行すると、全 Linux プロジェクトへ配備されます。影響範囲
Claude/templates/claude/START_PROMPT.mdのみ（既存スクリプトへの影響なし）テスト結果
check-doc-versions.js: PASSED (README=v3.2.90, CHANGELOG=v3.2.90)残課題
なし
🤖 Generated with Claude Code
Summary by CodeRabbit
New Features
Documentation
Bug Fixes