feat(goal): /goal に DynamicWorkflows + Verify 必須ゲート(CodeRabbit/Codex/security) を明記 - #300
Merged
Merged
Conversation
Windows/Linux 両 CLI が v2.1.156 native で dynamic workflows 利用可と実測確定したため、 CTO が適切な場面で実際に活用するよう /goal 標準指示に組み込む (ユーザー選択: 強め)。 - START_PROMPT.md の /goal 行: 「大規模監査/横断調査/多観点設計で token<70% かつ 残≥60min なら DynamicWorkflows(deep-research 等) を積極活用・ultracode 既定化禁止」を追記 - mis-trigger 回避: DynamicWorkflows(1語) + slash なし deep-research で記述 - verify-goal-set.js: 必須キーワードに DynamicWorkflows 追加 (8→9)、カウント表示を動的化 検証: verify-goal-set.js が「必須キーワード 9/9 整合」を確認。START_PROMPT は launcher 経由で 全プロジェクトに中央配布される (Linux git pull で反映)。 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
Note
|
| Layer / File(s) | Summary |
|---|---|
先頭 /goal 指示文への条件追加 Claude/templates/claude/START_PROMPT.md |
/goal 指示文に、DynamicWorkflows（deep-research等）を token 残量70%未満かつ残り60分以上の条件下で積極活用する指示と「ultracode既定化は禁止」を追記しました。 |
Verify Mandatory 文言の拡張 Claude/templates/claude/START_PROMPT.md |
Verify Mandatory 行を更新し、CodeRabbit review / Codex review（利用可能時）/ security scan（gitleaks・secret・npm audit）を必須実施として明記し、ultrareview（Gate-2b）は自律ループで自動実行しない旨を追記しました。 |
Estimated code review effort
🎯 2 (Simple) | ⏱️ ~10 minutes
Possibly related PRs
- Kensan196948G/ClaudeCode-StartUpTools-New#279: START_PROMPT.md の
/goalテンプレート構造変更と制約調整で重複する箇所があります。 - Kensan196948G/ClaudeCode-StartUpTools-New#281: ClaudeOS ブートローダー／
/goal指示文のリファクタで関連する起動指示の変更が含まれます。 - Kensan196948G/ClaudeCode-StartUpTools-New#274: START_PROMPT.md の
/goal/検証ルール強化に関する変更が類似しています。
🚥 Pre-merge checks | ✅ 5
✅ Passed checks (5 passed)
| Check name | Status | Explanation |
|---|---|---|
| Description Check | ✅ Passed | Check skipped - CodeRabbit’s high-level summary is enabled. |
| Title check | ✅ Passed | PRタイトルは、/goal指示にDynamicWorkflowsと必須検証ゲート（CodeRabbit/Codex/security scan）を追加する変更を正確に要約しており、主要な変更内容を明確に反映しています。 |
| Docstring Coverage | ✅ Passed | No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. |
| Linked Issues check | ✅ Passed | Check skipped because no linked issues were found for this pull request. |
| Out of Scope Changes check | ✅ Passed | Check skipped because no linked issues were found for this pull request. |
✏️ Tip: You can configure your own custom pre-merge checks in the settings.
✨ Finishing Touches
🧪 Generate unit tests (beta)
- Create PR with unit tests
- Commit unit tests in branch
feat/goal-dynamic-workflows
Comment @coderabbitai help to get the list of available commands and usage tips.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
Claude/templates/claude/START_PROMPT.md (1)
10-11:⚠️ Potential issue | 🟡 Minor | ⚡ Quick win必須キーワード数の説明が古いです。
Line 11 の「
8 個整合」は、今回の8→9変更と不整合です。固定値を避けて「必須キーワード整合をチェック」のような動的表現に更新してください。🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@Claude/templates/claude/START_PROMPT.md` around lines 10 - 11, Update the hardcoded "8 個整合" wording in the SessionStart description to a dynamic phrasing (e.g., "必須キーワード整合をチェック") so it no longer references a specific count; edit the START_PROMPT.md text that describes the SessionStart hook (verify-goal-set.js) and replace the exact token "8 個整合" with a neutral/dynamic expression that conveys checking required keyword alignment.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@Claude/templates/claude/START_PROMPT.md`:
- Line 1: The prompt text currently says "token残量70%未満" but the runtime gate
checks tokenUsedPct < 70 (usage percentage), so update the START_PROMPT.md
phrase to match the implementation by changing "token残量70%未満" to
"token使用率70%未満"; verify consistency with the check in session-start.js (the
tokenUsedPct < 70 condition) so wording and logic align.
---
Outside diff comments:
In `@Claude/templates/claude/START_PROMPT.md`:
- Around line 10-11: Update the hardcoded "8 個整合" wording in the SessionStart
description to a dynamic phrasing (e.g., "必須キーワード整合をチェック") so it no longer
references a specific count; edit the START_PROMPT.md text that describes the
SessionStart hook (verify-goal-set.js) and replace the exact token "8 個整合" with
a neutral/dynamic expression that conveys checking required keyword alignment.
🪄 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: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 46df6200-37f1-4f55-ac54-b51c57419cfc
⛔ Files ignored due to path filters (1)
.claude/claudeos/scripts/hooks/verify-goal-set.jsis excluded by none and included by none
📒 Files selected for processing (1)
Claude/templates/claude/START_PROMPT.md
CodeRabbit 指摘 (Major): /goal の「token残量70%未満」が session-start.js の tokenUsedPct < 70 (使用率70%未満) と逆向きだった。実装・core/04「token消費<70%」に 合わせて「token使用率70%未満」に統一。 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…上位原則に明記 ユーザー要望: Codex review / CodeRabbit review / security scan を START_PROMPT に格上げ。 core §8/§8.5/§11 + STABLE §9 で既に指示済みだが、ブート指示に明記して強制力を上げる。 - /goal 行: Verify で CodeRabbit review・Codex review(利用可能時)・security scan (gitleaks/secret/npm audit) を必ず実施し STABLE 判定の前提とする旨を追記 - 最上位原則: Verify Mandatory を具体化 (3 ゲート + 参照 + ultrareview は自律実行しない注記) - Codex は §0 に従い「利用可能時」(任意)。ultrareview(Gate-2b) は課金・人手起動のため除外 - verify-goal-set.js: 必須キーワードに CodeRabbit / security 追加 (9→11) 検証: verify-goal-set.js が「必須キーワード 11/11 整合」を確認。 Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Kensan196948G
changed the title
Owner
Author
追加コミット
|
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
1 participant
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.
📌 背景
Windows/Linux 両 CLI が v2.1.156 native で dynamic workflows が利用可能と実測確定。機能は「存在する」が、CTO が実際に活用するには /goal 標準指示への組み込みが効果的（ユーザー選択: 強め）。
🛠 変更
START_PROMPT.mdの/goal行に、AgentTeams と同格で DynamicWorkflows 活用指針を追記:DynamicWorkflows(1語・bare "workflow" を避ける) + slash なしdeep-research（コマンド誤認防止）。/goal は UI コマンド引数なのでブート時 workflow 化はしない設計verify-goal-set.js: 必須キーワードにDynamicWorkflows追加（8→9）、カウント表示を動的化✅ 検証
配布
START_PROMPT.mdは launcher (Start-ClaudeCode.ps1 / Linux cron) が中央読込し全プロジェクトに渡すため、merge + Linuxgit pullで全プロジェクトに反映。per-project 配布不要。影響範囲
/goal行への 1 文追記 + キーワード 1 個追加のみ。既存 8 キーワード・他指示は不変。🤖 Generated with Claude Code
Summary by CodeRabbit