Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Slackの同期データからメンションを検索しTODO/タスクを抽出・ステータス判定するスキル。 「Slackからタスク抽出」「TODO確認」「メンション確認」等のリクエストで発動。
.claude/skills/minicoohei-slack-todo-extractor/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -55% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -38% | 0% |
Slackの同期データ(slack-sync/data/)から、特定ユーザー宛のメンションを検索し、スレッド返信を含めてTODO/タスクを抽出・ステータス判定するスキルです。
bash# 基本(キーワードベース) uv run python skills/slack-todo-extractor/scripts/extract_todos.py \ --users "YourName,your-username" \ --period "2026-01-06:2026-01-08" # LLMベース(高精度、要GEMINI_API_KEY) uv run python skills/slack-todo-extractor/scripts/extract_todos.py \ --users "YourName,your-username" \ --period "1/6:8" \ --use-llm
| パラメータ | 必須 | 説明 | 例 | |-----------|------|------|-----| | --users, -u | Yes | 対象ユーザー名(カンマ区切り) | YourName, your-username | | --period, -p | Yes | 検索期間 | 2026-01-06:2026-01-08 or 1/6:8 | | --workspace, -w | No | ワークスペース(省略時は全て) | my-workspace, my-workspace-2 | | --use-llm | No | LLM(Gemini 2.0 Flash)で判定 | - | | --output, -o | No | 出力形式 | markdown(デフォルト)or json |
slack-sync/data/{workspace}/*.md から @ユーザー名 を含むメッセージを検索
各メンションに対して:
> #### 形式)を抽出--use-llm なし)| 条件 | ステータス | |------|-----------| | 対象ユーザーが「完了」「対応しました」等 | completed | | 依頼者が「ありがとう」「確認します」等 | completed | | 対象ユーザーが「承知」「やります」等 | in_progress | | 返信なし | pending |
--use-llm あり)Gemini 2.0 Flashが文脈を理解して判定:
Credential Store に設定:
bashuv run python tools/credential_manager.py store GEMINI_API_KEY
slack-sync/ のSlack同期が完了していること(data/slack-sync/ 参照)slack-search: Slackメッセージの全文検索slack-task-manager: 統合タスク管理| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 14,183 | 7,424 | -48% | 1 | 1 | 0% | 2,557 | 1,151 | -55% | 0 | 0 | — |
case-02 | fail→fail | 11,108 | 3,743 | -66% | 1 | 1 | 0% | 1,823 | 1,189 | -35% | 0 | 0 | — |
case-21 | pass→fail | 12,797 | 16,423 | +28% | 1 | 1 | 0% | 2,100 | 3,920 | +87% | 0 | 0 | — |
case-03 | fail→fail | 13,501 | 6,036 | -55% | 1 | 1 | 0% | 2,395 | 1,155 | -52% | 0 | 0 | — |
case-04 | fail→pass | 15,560 | 3,232 | -79% | 1 | 1 | 0% | 2,899 | 1,315 | -55% | 0 | 0 | — |
case-05 | fail→fail | 4,027 | 3,318 | -18% | 1 | 1 | 0% | 528 | 1,158 | +119% | 0 | 0 | — |
case-06 | fail→fail | 12,221 | 4,495 | -63% | 1 | 1 | 0% | 2,196 | 1,640 | -25% | 0 | 0 | — |
case-22 | fail→fail | 16,252 | 3,749 | -77% | 1 | 1 | 0% | 3,092 | 923 | -70% | 0 | 0 | — |
case-07 | fail→pass | 17,711 | 2,806 | -84% | 1 | 1 | 0% | 3,231 | 1,221 | -62% | 0 | 0 | — |
case-08 | fail→pass | 9,913 | 6,461 | -35% | 1 | 1 | 0% | 1,735 | 1,395 | -20% | 0 | 0 | — |
case-09 | fail→pass | 14,036 | 3,970 | -72% | 1 | 1 | 0% | 2,394 | 1,349 | -44% | 0 | 0 | — |
case-10 | pass→pass | 11,717 | 3,159 | -73% | 1 | 1 | 0% | 1,855 | 1,261 | -32% | 0 | 0 | — |
case-11 | pass→pass | 11,502 | 2,205 | -81% | 1 | 1 | 0% | 1,840 | 1,073 | -42% | 0 | 0 | — |
case-12 | fail→pass | 10,109 | 1,759 | -83% | 1 | 1 | 0% | 1,635 | 1,011 | -38% | 0 | 0 | — |
case-13 | pass→pass | 14,221 | 4,732 | -67% | 1 | 1 | 0% | 2,256 | 1,575 | -30% | 0 | 0 | — |
case-14 | pass→pass | 12,022 | 3,742 | -69% | 1 | 1 | 0% | 1,817 | 1,312 | -28% | 0 | 0 | — |
case-15 | fail→pass | 4,750 | 4,736 | -0% | 1 | 1 | 0% | 834 | 1,684 | +102% | 0 | 0 | — |
case-16 | fail→fail | 4,179 | 5,979 | +43% | 1 | 1 | 0% | 703 | 1,781 | +153% | 0 | 0 | — |
case-17 | fail→pass | 8,014 | 1,428 | -82% | 1 | 1 | 0% | 1,336 | 948 | -29% | 0 | 0 | — |
case-18 | fail→pass | 9,908 | 3,749 | -62% | 1 | 1 | 0% | 1,968 | 1,489 | -24% | 0 | 0 | — |
case-19 | pass→pass | 11,147 | 3,222 | -71% | 1 | 1 | 0% | 1,751 | 1,193 | -32% | 0 | 0 | — |
case-20 | pass→pass | 9,293 | 4,033 | -57% | 1 | 1 | 0% | 1,660 | 1,369 | -18% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +32 percentage points is the difference between those two pass rates over the 19 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.