菜单
Searches the web via DuckDuckGo and returns results as markdown, CSV, or JSON. Use this skill whenever the user wants to search the web, look up information online, find URLs, do web search, research a topic, or needs current/web-based answers.
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。
基于 SOC 职业分类
Phase/task based workflow system with PLAN.md as single source of truth. Use when tackling projects that require structured iteration through Planning, Analysis, Design, Implementation, Testing, Deployment, Maintenance, etc phases with clear dependency graphs.
Manages Python packages, projects, scripts, tools, and Python versions with extreme speed. Use when the user mentions uv, python packages, pip replacement, virtual environments, venv, pyproject.toml, dependency management, python version switching, tool installation (uvx), script dependencies, lockfiles, workspace management, or any Python packaging task. Replaces pip, pip-tools, pipx, poetry, pyenv, twine, and virtualenv.
jq 1.8.2 — lightweight command-line JSON processor. Use when the user needs to parse, query, transform, or manipulate JSON data from the command line, process API responses, extract fields from JSON, convert between formats (JSON-to-CSV, JSON-to-XML), validate JSON, or work with any structured data in JSON format. Covers filters, builtins, regex, modules, streaming, and all jq 1.8.2 features.
Operate rqlite 10.2.4 — distributed SQLite database with Raft consensus. Use for deploying, configuring, querying, backing up, and managing rqlite clusters. Covers rqlited server flags, HTTP API (/db/execute, /db/query, /db/request), CLI (.status, .backup, .nodes), clustering (join, DNS, Consul, etcd, Kubernetes), TLS/mTLS auth, queued writes, CDC, consistency levels, and SQLite extensions. Use whenever the user mentions rqlite, distributed SQLite, Raft database, or needs a lightweight fault-tolerant relational store.
Query, transform, and convert YAML, JSON, XML, INI, TOML, HCL, CSV, TSV, Properties, Lua, Shell variables, and Base64 data using jq-like expressions. Use when the user mentions yq, YAML processing, JSON-to-YAML conversion, config file manipulation, Kubernetes YAML editing, or any structured data transformation task. Supports multi-document YAML, anchors/aliases, comments, and in-place updates.
Ruff 0.15.18 — extremely fast Python linter and code formatter written in Rust. Use this skill whenever the user mentions ruff, linting Python, formatting Python code, ruff check, ruff format, pyproject.toml lint config, Flake8 replacement, Black replacement, isort replacement, or any Python code quality task involving ruff configuration, rule selection, or CI integration. Covers `ruff check`, `ruff format`, `ruff rule`, `ruff config`, `ruff analyze graph`, and `ruff server` (LSP).
| name | websearch |
| description | Searches the web via DuckDuckGo and returns results as markdown, CSV, or JSON. Use this skill whenever the user wants to search the web, look up information online, find URLs, do web search, research a topic, or needs current/web-based answers. |
| metadata | {"tags":["meta"]} |
Search the web via DuckDuckGo and return structured results.
websearch fetches search results from DuckDuckGo, parses the HTML response, and outputs them in a structured format.
Supports three output formats: markdown (default), CSV, and JSON. Results can be written to stdout or a file.
# Basic search (markdown to stdout)
websearch.sh "react hooks tutorial"
# JSON output to file
websearch.sh "python async patterns" --format json --output results.json
# CSV output
websearch.sh "rust vs go 2025" --format csv
# Help
websearch.sh --help
| Flag | Default | Description |
|---|---|---|
--format | markdown | Output format: markdown, csv, json |
--output | stdout | Output destination: stdout or a file path |
uddg parameter contains the actual URL (URL-encoded). The parser extracts and decodes it automatically.