據 AlphaMoat 統計,截至 2026-10,Parallel Deep Research 累計下載 5,570 次、獲 3 個星標,按累計下載量在本次參與排名的 10,914 個 AI 技能中排名第 5,182 位,在「知識管理」分類(745 個)中排名第 364 位。
列表預設按累計下載量排序,也可按星標排序;詳情頁排名按累計下載量計算。
| 平台 | 累計下載 | 星標 | 評論 |
|---|---|---|---|
| ClawHub ↗ | 5570 | 3 | - |
各平台的計數口徑不同,統計時間不一致。
用已安裝驗證的 parallel-cli 提交限定範圍的問題或檔案,選擇處理器並儲存 JSON 或 Markdown;也可非阻塞啟動後按任務 ID 查詢。適合深入綜合,快速查詢、即時新聞和特定 URL 擷取使用其他工具。安裝或驗證失敗應停下解決。問題與檔案傳送給外部 Parallel,須符合資料政策;報告保留磁碟,謹慎選儲存路徑。使用結論前檢查來源及衝突。
資料釋出於 2026-10-27
以下是根據技能文件生成的示例,不代表實際執行結果。
2026-09 起統計方法調整:同名技能按作者分別統計,並覆蓋全部技能。圖中虛線為調整前的月份,與之後的資料不直接比較,也不計算跨越調整的增長。
| 月份 | 累計下載 | 環比 | 星標 |
|---|---|---|---|
| 2026-10 | 5570 | — | 3 |
| 2026-09 | 5444 | — | 3 |
| 2026-08 | 5335 | +3.1% | 3 |
| 2026-07 | 5174 | +5.7% | 3 |
| 2026-06 | 4893 | +3.8% | 3 |
| 2026-05 | 4715 | — | 3 |
| 2026-03 | 3725 | — | 3 |
以下是作者釋出的原始檔案,按純文字顯示,本站未做修改或稽核;其中的連結不可點選,命令請核實後再執行。抓取於 2026-10-23。
--- name: parallel-deep-research description: "Deep multi-source research via Parallel API. Use when user explicitly asks for thorough research, comprehensive analysis, or investigation of a topic. For quick lookups or news, use parallel-search instead." homepage: https://parallel.ai --- # Parallel Deep Research Deep, multi-source research for complex topics requiring synthesis from many sources. Returns comprehensive reports with citations.
## When to Use
Trigger this skill when the user asks for:
- "deep research on...", "thorough investigation of...", "comprehensive report about..."
- "research everything about...", "full analysis of..."
- Complex topics requiring synthesis from 10+ sources
- Competitive analysis, market research, due diligence
- Questions where depth and accuracy matter more than speed
**NOT for:**
- Quick lookups or simple questions (use parallel-search)
- Current news or recent events (use parallel-search with `--after-date`)
- Reading specific URLs (use parallel-extract)
## Quick Start
```bash
parallel-cli research run "your research question" --processor pro-fast --json -o ./report
```
## CLI Reference
### Basic Usage
```bash
parallel-cli research run "<question>" [options]
```
### Common Flags
| Flag | Description |
|------|-------------|
| `-p, --processor <tier>` | Processor tier (see table below) |
| `--json` | Output as JSON |
| `-o, --output <path>` | Save results to file (creates .json and .md) |
| `-f, --input-file <path>` | Read query from file (for long questions) |
| `--timeout N` | Max wait time in seconds (default: 3600) |
| `--no-wait` | Return immediately, poll later with `research status` |
### Processor Tiers
| Processor | Time | Use Case |
|-----------|------|----------|
| `lite-fast` | 10-20s | Quick lookups |
| `base-fast` | 15-50s | Simple questions |
| `core-fast` | 15s-100s | Moderate research |
| `pro-fast` | 30s-5min | Exploratory research (default) |
| `ultra-fast` | 1-10min | Multi-source deep research |
| `ultra2x-fast` | 1-20min | Difficult deep research |
| `ultra4x-fast` | 1-40min | Very difficult research |
| `ultra8x-fast` | 1min-1hr | Most challenging research |
Non-fast variants (e.g., `pro`, `ultra`) take longer but use fresher data.
### Examples
**Basic research:**
```bash
parallel-cli research run "What are the latest developments in quantum computing?" \
--processor pro-fast \
--json -o ./quantum-report
```
**Deep competitive analysis:**
```bash
parallel-cli research run "Compare Stripe, Square, and Adyen payment platforms: features, pricing, market position, and developer experience" \
--processor ultra-fast \
--json -o ./payments-analysis
```
**Long research question from file:**
```bash
# Create question file
cat > /tmp/research-question.txt << 'EOF'
Investigate the current state of AI regulation globally:
1. What regulations exist in the US, EU, and China?
2. What's pending or proposed?
3. How do companies like OpenAI, Google, and Anthropic respond?
4. What industry groups are lobbying for/against regulation?
EOF
parallel-cli research run -f /tmp/research-question.txt \
--processor ultra-fast \
--json -o ./ai-regulation-report
```
**Non-blocking research:**
```bash
# Start research without waiting
parallel-cli research run "research question" --no-wait
# Check status later
parallel-cli research status <task-id>
# Poll until complete
parallel-cli research poll <task-id> --json -o ./report
```
## Best-Practice Prompting
### Research Question
Write 2-5 sentences describing:
- The specific question or topic
- Scope boundaries (time period, geography, industries)
- What aspects matter most (pricing? features? market share?)
- Desired output format (comparison table, timeline, pros/cons)
**Good:**
```
Compare the top 5 CRM platforms for B2B SaaS companies with 50-200 employees.
Focus on: pricing per seat, integration ecosystem, reporting capabilities.
Include recent 2024-2026 changes and customer reviews from G2/Capterra.
```
**Poor:**
```
Tell me about CRMs
```
## Response Format
Returns structured JSON with:
- `task_id` — unique identifier for polling
- `status` — `pending`, `running`, `completed`, `failed`
- `result` — when complete:
- `summary` — executive summary
- `findings[]` — detailed findings with sources
- `sources[]` — all referenced URLs with titles
## Output Handling
When presenting research results:
- Lead with the **executive summary** verbatim
- Present **key findings** without paraphrasing
- Include **source URLs** for all facts
- Note any **conflicting information** between sources
- Preserve all facts, names, numbers, dates, quotes
## Running Out of Context?
For long conversations, save results and use `sessions_spawn`:
```bash
parallel-cli research run "<question>" --json -o /tmp/research-<topic>
```
Then spawn a sub-agent:
```json
{
"tool": "sessions_spawn",
"task": "Read /tmp/research-<topic>.json and present the executive summary and key findings with sources.",
"label": "research-summary"
}
```
## Error Handling
| Exit Code | Meaning |
|-----------|---------|
| 0 | Success |
| 1 | Unexpected error (network, parse) |
| 2 | Invalid arguments |
| 3 | API error (non-2xx) |
## Prerequisites
Requires `parallel-cli` (installed and authenticated). If `parallel-cli --version` fails, or if a later command fails with an authentication error, tell the user to see https://docs.parallel.ai/integrations/cli and stop.
## References
- [API Docs](https://docs.parallel.ai)
- [Research API Reference](https://docs.parallel.ai/api-reference/research)
透過公開搜尋URL在16個搜尋引擎(7個國內直連+9個國際)中做跨引擎檢索、結果去重與多源交叉驗證,無需API Key,內建零依賴Python指令碼。Use when 使用者要求“搜一下/查資料/多找幾個來源/核實訊息/找公眾號文章/排查報錯/找論文或開源專案”時。輸入自然語言查詢詞,輸出帶來源連結與可信度標註的結構化結果。不適用於私有資料檢索、批次抓取、需要登入後可見的內容。
透過 IMA OpenAPI 管理筆記與知識庫內容。
把長文字、文件與網頁內容壓縮成要點和關鍵詞的摘要工具。
透過“查一下、搜一下、搜尋、找一下”等自然語言檢索公開網頁,整理來源並交付橫版 PDF、Markdown 與標題連結表。支援使用者主動開啟、檢視和取消技能收藏,保留原報告取回。
規定如何為網頁請求選擇合適工具的必讀指引。
AI HOT(aihot.virxact.com)中文 AI 資訊查詢 Skill,幫助使用者獲取今日 AI 圈動態、AI 日報、AI 熱點、最新 AI 資訊以及 OpenAI、Anthropic、Google 等公司的最新發布資訊。
全球12億文獻知識庫(8千萬中文期刊可下載)——透過 SmartLib 開放平台 API 提供中外文學術文獻檢索與下載能力,覆蓋 8000 萬篇授權中文期刊全文 + 12.28 億條全球文獻後設資料(期刊 7.19 億 / 專利 2.15 億 / 會議 7155 萬 / 學位論文 2473 萬 / 標準 268 萬)。 能力:中英文關鍵詞檢索、文獻詳情、中文期刊 PDF 全文下載、外文 OA 文獻十級渠道免費下載(不消耗配額)、智慧關鍵詞擴充套件、核心期刊優先排序、相關性重排、引文追溯、分類號檢索。 配額:首次使用自動註冊,免費 100 次檢索 + 10 次下載 / 月;耗盡自動彈出方案(體驗卡 / 個人版月 / 專業版月 / 單篇下載 / 下載包),可說「升級 / 充值」喚起;企業 / 機構定製聯絡 [email protected]。 觸發:使用者表達"查論文""找文獻""檢索學術""搜期刊""查專利""找標準""下論文""寫文獻綜述""找參考文獻""查 SCI/EI"等意圖時啟用;也適用"幫我找關於 XX 的論文""寫文獻綜述""找幾篇引用支撐論點"。英文:"find papers" "search literature" "write literature review" "find supporting citations"。 呼叫前必須先用 /consume 獲取 consume_token,再憑 token 調 /search(每次計費介面呼叫都需一次 /consume)。
用 uvx 直接呼叫 markitdown,把各類檔案轉成 Markdown。
資料月份:2026-10 · 下載/星標數聚合自公開技能註冊庫(ClawHub、SkillHub),按月更新。口徑說明見資料方法。