feat(local-llm): P-0 classifier opt-in + §10 ブランチ分離運用 (ADR-038 Phase 5)#123
Merged
feat(local-llm): P-0 classifier opt-in + §10 ブランチ分離運用 (ADR-038 Phase 5)#123
Conversation
§A-2 Phase 5 dogfood の前提セットアップ。pr-monitor-config.toml に
[classifier] section を追加し enabled=true でローカル LLM (mistral:7b) による
CodeRabbit findings classification を master 上で有効化。
§10 ブランチ分離運用を以下のように修正 (本セッション 2 ターン目のユーザー
directive 「P-1〜P-5 は PR を個別に実装する想定」に従い、過剰隔離を解消):
- §10.3: P-0 / P-1〜P-5 はすべて master ベース個別 PR (元案では feature 枝のみ
だったが、master の classifier は smoke test 動作確認済 + revert PR が即時
fallback として機能するため、過剰隔離だった)
- §10.4: dogfood は master 上で classifier 自動起動。枝切替不要
- §10.7: 再開チェックリストを A (P-N PR 着手) / B (§8.D/E/F 追加実装) に分割
- 隔離スコープを §8.D / §8.E / §8.F (新規実装) に限定
- §10.6 C/D の参照誤りを修正 (kill-switch は C、6 ヶ月経過は D)
§A-2 計測ログに P-0 完了状態を記録 (smoke test 結果含む)。Setup 手順から
枝 checkout step を削除 (master 上で完結する dogfood に対応)。
P-0 smoke test:
- Ollama: mistral:7b (4.0GB) + llama2:13b (6.9GB) 起動確認
- classifier exe (.claude/cli-finding-classifier.exe, 2.2MB) 配備確認
- finding {Major / unused import} → action=auto_fix / confidence=1.0 (妥当)
- cli-pr-monitor は新 [classifier] section を読み込み可、compile 通過
本 PR は feature/local-llm-dogfood → master の squash merge を想定。
📝 Walkthrough概要ローカルLLM分類器（Ollama）の段階的検証と統合のための構成スキーマを追加し、フェーズ5計測計画、マスター/フィーチャーブランチ間の分離戦略、実行ガイダンス、および環境チェックリストを定義します。 変更内容ローカルLLM検証統合ワークフロー
推定レビュー工数🎯 2 (Simple) | ⏱️ ~12 分 関連の可能性があるプルリクエスト
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (1)
docs/local-llm-offload-analysis.md (1)
354-356:⚠️ Potential issue | 🟡 Minor | ⚡ Quick winjq フィルタに余分な
)があり、コンパイルエラーになります。
test("[a-zA-Z]{8,}")の後の)が対応する(なしに現れており、実行するとjq: error: compile errorで失敗します。🐛 修正案
-jq '[.classified_findings[] | select(.normalized_issue) | .normalized_issue | test("[a-zA-Z]{8,}"))]' .claude/pr-monitor-state.json +jq '[.classified_findings[] | select(.normalized_issue) | .normalized_issue | test("[a-zA-Z]{8,}")]' .claude/pr-monitor-state.jsonなお、このコマンドは
true/falseのブール値配列を返します。違反件数を直接カウントする場合は末尾に| lengthを追加するか、select(test(...))形式に変更するとより直接的です。🤖 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 `@docs/local-llm-offload-analysis.md` around lines 354 - 356, The jq expression contains an extra closing parenthesis after test("[a-zA-Z]{8,}") causing a compile error; remove the stray ')' so the expression becomes valid and, if you want a count rather than a boolean array, append | length or rewrite the pipeline to use select(test(...)) to return matching objects directly; target the jq command line shown (the expression using .classified_findings[], .normalized_issue, and test("[a-zA-Z]{8,}")) and fix the unmatched parenthesis and optionally add | length or change to select(test(...)) as described.
🤖 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.
Outside diff comments:
In `@docs/local-llm-offload-analysis.md`:
- Around line 354-356: The jq expression contains an extra closing parenthesis
after test("[a-zA-Z]{8,}") causing a compile error; remove the stray ')' so the
expression becomes valid and, if you want a count rather than a boolean array,
append | length or rewrite the pipeline to use select(test(...)) to return
matching objects directly; target the jq command line shown (the expression
using .classified_findings[], .normalized_issue, and test("[a-zA-Z]{8,}")) and
fix the unmatched parenthesis and optionally add | length or change to
select(test(...)) as described.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 2f97307c-a4da-4cb9-8e38-0e8122710b05
📒 Files selected for processing (2)
docs/local-llm-offload-analysis.mdpr-monitor-config.toml
6 tasks
aloekun
added a commit
that referenced
this pull request
May 7, 2026
This was referenced May 7, 2026
Merged
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.
Summary
docs/local-llm-offload-analysis.mdに §10「ブランチ分離運用」を追加し、P-0 (config opt-in) としてpr-monitor-config.tomlに[classifier] enabled=trueを landfeature/local-llm-dogfood枝に残置 — 検証完了まで master に流さないChanges
docs/local-llm-offload-analysis.md__backup-claude-config/手動運用、強制 hook 化なし)pr-monitor-config.toml[classifier]section 追加 (enabled=true/model=mistral:7b/endpoint=http://localhost:11434/timeout_secs=30)enabled=falseに戻す)P-0 動作確認
mistral:7b(4.0GB) +llama2:13b(6.9GB) 起動確認.claude/cli-finding-classifier.exe, 2.2MB) 配備確認{Major / unused import / remove statement}→action=auto_fix/action_confidence=1.0(妥当な分類)cli-pr-monitorは新[classifier]section を読み込み可、compile 通過背景 — モデル簡素化の経緯
本セッション 1 ターン目で「ローカル LLM の検証作業を feature ブランチに分離したい」というユーザー要望に応じ、§10 で Option B (隔離ベース) を策定。当初は P-0 / P-1〜P-5 も feature 枝のみとしていたが、2 ターン目で「P-1〜P-5 は PR を個別に実装する想定」というユーザー directive を受け、隔離スコープを §8.D / §8.E / §8.F (新規実装) に限定 する model に整理。
判断根拠:
kill-switch
classifier 起動が予期せぬ問題を起こした場合は revert PR で
[classifier] enabled = falseに戻す のが kill-switch (§10.6 C 簡易版)。crate 削除等の物理削除は dogfood 失敗判定後にまとめて実施 (§10.6 C 完全版)。Test plan
curl -s http://localhost:11434/api/tags)ls -la .claude/cli-finding-classifier.exe)echo '[...]' | .claude/cli-finding-classifier.exe --timeout-secs 60→ JSON 妥当)cargo run --bin cli-pr-monitorで no parse error)関連
Summary by CodeRabbit
リリースノート
ドキュメント
新機能