Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Thread Manager 为 Claude Code 提供多线程对话管理功能，让你可以并行处理多个任务，保持上下文隔离和专注。
.claude/skills/peterfei-thread-manager-skill/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -51% | 0% |
Thread Manager 为 Claude Code 提供多线程对话管理功能，让你可以并行处理多个任务，保持上下文隔离和专注。
本 Skill 通常作为 ai-agent-team 项目的一部分安装。如果需要单独安装：
npm install 安装依赖。npm run build 编译代码。config.json) 中。| 工具名称 | 描述 | |---------|------| | create_thread | 创建新线程 | | list_threads | 列出所有线程 | | switch_thread | 切换到指定线程 | | get_thread | 获取线程详情 | | update_thread | 更新线程信息 | | delete_thread | 删除线程 | | get_current_thread | 获取当前活跃线程 | | track_file_change | 记录文件变更 (支持自动检测) | | search_messages | 基于语义搜索历史消息 (支持自然语言) |
/thread new "Title": 创建新线程/threads: 列出所有线程/thread <id>: 切换线程/thread info: 查看当前线程详情/memory search "query": 搜索历史消息 (需要调用 search_messages)数据存储在 ~/.claude/threads/threads.db (默认) 或环境变量 THREADS_DB_PATH 指定的位置。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-09 | fail→pass | 11,576 | 2,434 | -79% | 1 | 1 | 0% | 1,849 | 758 | -59% | 0 | 0 | — |
case-02 | fail→fail | 3,804 | 13,587 | +257% | 1 | 1 | 0% | 499 | 2,195 | +340% | 0 | 0 | — |
case-01 | fail→fail | 4,454 | 5,353 | +20% | 1 | 1 | 0% | 639 | 736 | +15% | 0 | 0 | — |
case-03 | fail→fail | 4,351 | 2,384 | -45% | 1 | 1 | 0% | 611 | 794 | +30% | 0 | 0 | — |
case-04 | fail→pass | 8,567 | 2,578 | -70% | 1 | 1 | 0% | 1,429 | 873 | -39% | 0 | 0 | — |
case-05 | fail→pass | 9,280 | 3,268 | -65% | 1 | 1 | 0% | 1,406 | 675 | -52% | 0 | 0 | — |
case-06 | pass→pass | 6,110 | 1,186 | -81% | 1 | 1 | 0% | 890 | 567 | -36% | 0 | 0 | — |
case-07 | pass→pass | 5,819 | 1,581 | -73% | 1 | 1 | 0% | 897 | 701 | -22% | 0 | 0 | — |
case-08 | pass→pass | 9,536 | 1,947 | -80% | 1 | 1 | 0% | 1,500 | 694 | -54% | 0 | 0 | — |
case-10 | fail→pass | 10,076 | 1,559 | -85% | 1 | 1 | 0% | 1,557 | 642 | -59% | 0 | 0 | — |
case-11 | fail→pass | 7,704 | 1,626 | -79% | 1 | 1 | 0% | 1,220 | 592 | -51% | 0 | 0 | — |
case-12 | fail→pass | 10,588 | 1,779 | -83% | 1 | 1 | 0% | 1,487 | 703 | -53% | 0 | 0 | — |
case-13 | fail→pass | 12,190 | 2,084 | -83% | 1 | 1 | 0% | 1,852 | 708 | -62% | 0 | 0 | — |
case-14 | pass→pass | 11,928 | 1,337 | -89% | 1 | 1 | 0% | 1,758 | 627 | -64% | 0 | 0 | — |
case-15 | fail→pass | 7,092 | 1,773 | -75% | 1 | 1 | 0% | 990 | 704 | -29% | 0 | 0 | — |
case-16 | pass→pass | 5,780 | 1,684 | -71% | 1 | 1 | 0% | 921 | 690 | -25% | 0 | 0 | — |
case-17 | fail→pass | 8,091 | 5,606 | -31% | 1 | 1 | 0% | 1,519 | 1,388 | -9% | 0 | 0 | — |
case-18 | pass→pass | 14,556 | 8,357 | -43% | 1 | 1 | 0% | 2,322 | 1,701 | -27% | 0 | 0 | — |
case-19 | fail→pass | 7,212 | 1,766 | -76% | 1 | 1 | 0% | 1,058 | 685 | -35% | 0 | 0 | — |
case-20 | fail→fail | 13,867 | 10,484 | -24% | 1 | 1 | 0% | 2,536 | 2,507 | -1% | 0 | 0 | — |
case-21 | fail→fail | 8,397 | 5,248 | -38% | 1 | 1 | 0% | 1,443 | 1,406 | -3% | 0 | 0 | — |
case-22 | fail→fail | 13,003 | 11,108 | -15% | 1 | 1 | 0% | 2,236 | 2,382 | +7% | 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 21 counted toward the lift figure. The other 1 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 +45 percentage points is the difference between those two pass rates over the 21 comparable cases.
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.