菜单
This skill should be used when the user asks about "firecrawl CLI", "jina CLI", "firecrawl command line", "jina command line", "scrape from terminal", "search from shell", or needs ready-to-use CLI commands for web scraping and search.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Agents-Store/claude-plugins --skill cli-recipes命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
This skill should be used when the user reports a problem with macstack-dev or the MACSTACK standard — "this skill did the wrong thing", "the schema is missing a field", "the passport for X is wrong", "send macstack feedback", "improve the schema based on my edit", "fix the macstack plugin" — and the fix must land in the right source (plugin, schema repo, or registry repo).
This skill should be used when the user asks to "set up Infisical for this project", "create .infisical.json", "pull the env keys", "wire the env", "sync secrets", or scaffold-project reaches the env step. Creates .infisical.json, pulls .env.prod/.env.dev, ensures every key from macstack.json resources.accesses exists, and installs the mandatory secrets scripts and commands.
This skill should be used when the user asks to "scaffold the project from macstack.json", "create the project working files", "generate project files from the spec", "build the project from macstack.json", or after a macstack.json is validated and the working tree must be built. Creates project files strictly in the prototype → stack plugins → dev plugins order.
基于 SOC 职业分类
| name | cli-recipes |
| description | This skill should be used when the user asks about "firecrawl CLI", "jina CLI", "firecrawl command line", "jina command line", "scrape from terminal", "search from shell", or needs ready-to-use CLI commands for web scraping and search. |
CLI commands for Firecrawl and Jina — the two services with official CLIs.
npm install -g firecrawl-cli
# or run without install:
npx -y firecrawl-cli@latest
firecrawl login --browser # Browser-based login
firecrawl login --api-key fc-YOUR_KEY # Direct key
firecrawl view-config # Show current auth/config
firecrawl logout # Sign out
firecrawl https://example.com/page
firecrawl scrape https://example.com/page --format markdown,links
firecrawl scrape https://example.com/spa --wait-for 3000 # JS rendering
firecrawl scrape https://example.com --screenshot # Capture screenshot
firecrawl scrape https://example.com -o output.md # Save to file
firecrawl scrape https://example.com --json --pretty # JSON output
firecrawl search "Next.js server components tutorial"
firecrawl search "React hooks" --limit 20
firecrawl search "TypeScript patterns" --scrape # Search + scrape results
firecrawl search "AI news" --tbs qdr:d # Last 24 hours
firecrawl search "tech startups" --location "San Francisco"
firecrawl crawl https://docs.example.com --wait --progress
firecrawl crawl https://docs.example.com --limit 50 --max-depth 3
firecrawl crawl https://docs.example.com --include-paths "/docs/*"
firecrawl crawl https://docs.example.com --exclude-paths "/blog/*"
firecrawl crawl JOB_ID # Check existing crawl status
firecrawl map https://example.com
firecrawl map https://example.com --limit 200
firecrawl map https://example.com --allow-subdomains
firecrawl agent "Find top 5 headless CMS and compare pricing"
firecrawl agent "Research React state management libraries" --wait
firecrawl agent "Find API rate limits" --urls https://docs.stripe.com --schema schema.json
Drive dynamic pages with natural language or code — replaces the old firecrawl browser session flow:
firecrawl interact https://example.com --prompt "log in and open the dashboard"
Recurring scrapes with change detection:
firecrawl monitor create https://example.com/pricing
firecrawl monitor list
Search GitHub issues, PRs, READMEs, and docs:
firecrawl developer "nextjs hydration mismatch"
| Flag | Description |
|---|---|
--json | Output as JSON |
--pretty | Pretty-print JSON |
-o FILE | Save output to file |
--timing | Show request timing |
--only-main-content | Strip boilerplate |
--api-key KEY | Override API key |
pip install jina-cli
# or
uv pip install jina-cli
export JINA_API_KEY=jina_xxx
jina read https://example.com/page
jina read https://example.com --json # Structured output
jina search "React hooks best practices"
jina search "React hooks" -n 10 # Number of results
jina search "AI news" --time d # Past day (d/w/m)
jina search "tech" --gl us --hl en # Geo/language targeting
jina search "transformer models" --arxiv # Search arXiv
jina search "financial analysis" --ssrn # Search SSRN
jina search "UI components" --images # Image search
jina search "Jina releases" --blog # Jina blog
jina embed "text to embed"
jina rerank "query" < documents.txt
jina classify "product review text" --labels positive,negative,neutral
jina dedup < urls.txt # Deduplicate
jina screenshot https://example.com -o screenshot.png
jina screenshot https://example.com --full-page
jina bibtex "attention is all you need"
jina expand "machine learning optimization"
jina pdf https://arxiv.org/pdf/2301.00001 --type figure,table
jina datetime https://example.com/article # Guess publish date
Jina CLI follows Unix philosophy — pipe commands together:
# Search and rerank
jina search "transformer models" | jina rerank "efficient inference"
# Read multiple URLs
cat urls.txt | jina read
# Search and deduplicate
jina search "attention mechanism" | jina dedup
# Search, read, and save
jina search "React patterns" -n 5 | jina read > research.md
Run embed, rerank, classify, and dedup locally without API key:
pip install jina-grep
jina grep serve start # Start local embedding server
jina embed --local "hello world"
jina classify --local "text" --labels positive,negative
Caveat: jina-grep is not currently published on PyPI, so pip install jina-grep fails as of 2026-08 (upstream jina-cli docs give this same instruction; an npm package named jina-grep exists but is an empty placeholder). Check https://github.com/jina-ai/jina-grep-cli for the actual install method before relying on local mode.
| Flag | Description |
|---|---|
--json | Structured JSON output |
-n NUM | Number of results |
--time d/w/m | Time filter (day/week/month) |
--gl CODE | Geographic location |
--local | Run locally (Apple Silicon) |
# Use Firecrawl to find URLs, Jina to process
firecrawl map https://docs.example.com --json | \
jq -r '.urls[]' | \
head -10 | \
while read url; do jina read "$url"; done > all_docs.md
firecrawl search "React server components" --json > firecrawl_results.json
jina search "React server components" --json > jina_results.json