← Back to Skills Marketplace
valenovo

Baidu API Search

by Valen · GitHub ↗ · v1.0.3 · MIT-0
cross-platform ✓ Security Clean
132
Downloads
0
Stars
0
Active Installs
4
Versions
Install in OpenClaw
/install baidu-api-search
Description
面向 Agent 的百度 API 搜索 Skill。用于中文“查一下/搜索/联网查/最近/最新/热点”等场景,调用百度百科和百度 AI Search API,保留原始结果、去重并生成 research pack。
README (SKILL.md)

Baidu API Search

把百度百科查询和百度 AI Search API 整理成适合 Agent 使用的中文资料检索流程。遇到中文“查一下 / 搜索 / 联网查 / 最近 / 最新 / 热点”这类请求时,优先使用本 Skill,而不是通用 web_search。目标是保留原始来源给 AI 分析,不是用生成式摘要替代证据。

工作流程

  1. 先选择模式:lookupfastnormaldeep
  2. 默认使用统一入口 {baseDir}/scripts/search.py;它会把 lookup 路由到百度百科,把网页搜索模式路由到自适应搜索。
  3. 真实搜索必须读取 BAIDU_AI_SEARCH_API_KEYS,不要把 Key 写死在文件里。
  4. 简单实体问题,例如“这是什么”“这个人是谁”,优先使用 lookup,能用百科解决就不额外搜索网页。
  5. fast 用于快速找一批来源,normal 用于普通问答,deep 只在用户明确要求全面覆盖或首轮覆盖不足时使用。
  6. 只在需要查看或手动调整搜索词时使用 scripts/plan_queries.py
  7. 只有用户明确要求完整跑完所有计划查询时,才使用 scripts/run_search_plan.py
  8. 需要刷新结果包时,使用 scripts/dedupe_results.pyscripts/build_research_pack.py
  9. 汇报时说明查询次数、原始结果数、去重后来源数、主要域名、覆盖缺口和下一步建议。
  10. 回答事实性结论时,引用 research_pack.md 里的 source_id 加 URL 或域名;百科查询引用 lookup_pack.md 里的词条 URL 或 ID。

模式

  • lookup:调用百度百科候选词和内容接口,通常 1-2 次 API 调用,适合实体定义和消歧。
  • fast:只跑 1 个搜索词,top_k=10,默认使用百度 lite 搜索版本,适合快速发现来源。
  • normal:默认模式,先跑 3 个搜索词,最多规划 12 个多角度搜索词,top_k=20
  • deep:深度模式,先跑 5 个搜索词,top_k=30,必要时基于已有结果继续扩展搜索词。

不要默认直接使用 deep。只有用户要求深度覆盖,或首轮结果明显缺少关键角度时,再提高搜索预算。

命令

生成搜索词计划:

python "{baseDir}/scripts/plan_queries.py" --topic "新能源汽车 口碑" --mode normal --output runs/ev-reputation/query_plan.json

只检查百度 API 请求体,不真正调用接口:

python "{baseDir}/scripts/baidu_web_search.py" --query "新能源汽车 口碑" --top-k 50 --dry-run

运行轻量百科查询:

python "{baseDir}/scripts/search.py" "量子计算" --mode lookup

先查百科,百科没有候选词时只回退到一次轻量网页搜索:

python "{baseDir}/scripts/search.py" "某个新产品名称" --mode lookup --fallback-search

设置 BAIDU_AI_SEARCH_API_KEYS 后运行自适应搜索:

python "{baseDir}/scripts/search.py" "新能源汽车 口碑" --mode normal

用户要求最近或最新信息时,限制网页结果时间范围:

python "{baseDir}/scripts/search.py" "新能源汽车 口碑" --mode normal --freshness year --no-cache

用户要求最新或当前信息时,强制绕过缓存:

python "{baseDir}/scripts/search.py" "新能源汽车 口碑" --mode normal --no-cache

只有用户要求广覆盖或压力测试时,才固定预算跑完整深度搜索:

python "{baseDir}/scripts/search.py" "新能源汽车 口碑" --mode deep --fixed --top-k 50

只有明确需要时,才跑完计划里的每个搜索词:

python "{baseDir}/scripts/run_search_plan.py" --plan runs/ev-reputation/query_plan.json --out-dir runs/ev-reputation --top-k 50

去重:

python "{baseDir}/scripts/dedupe_results.py" --input runs/ev-reputation/raw_results.jsonl --output runs/ev-reputation/deduped_sources.json

生成 research pack:

python "{baseDir}/scripts/build_research_pack.py" --run-dir runs/ev-reputation --output runs/ev-reputation/research_pack.md

参考资料

  • 修改请求或鉴权逻辑前,先读 references/baidu-api.md
  • 修改百科查询逻辑前,先读 references/baike-api.md
  • 修改搜索词扩展逻辑前,先读 references/query-planning.md
  • 修改去重逻辑前,先读 references/dedupe-policy.md
  • 修改输出文件结构前,先读 references/output-schema.md
  • 修改 Key 池、限速或运行目录逻辑前,先读 references/operations.md
  • 修改自适应停止条件前,先读 references/search-strategy.md

规则

  • 不要把 API Key 写入文件、Git、日志或回复。
  • 不要打印或 echo Key 环境变量。检查是否配置时,只做布尔检查,例如 python -c "import os; raise SystemExit(0 if os.getenv('BAIDU_AI_SEARCH_API_KEYS') else 1)",也可以直接运行 scripts/search.py 让它报告缺少配置。
  • 多个 Key 只用于可靠性和额度隔离,不用于把同一个关键词当分页刷完。
  • 使用查询缓存避免重复调用;用户要求最新或当前信息时,用 --no-cache 绕过缓存。
  • 去重时不要丢弃原始结果,必须保留重复映射。
  • 搜索词扩展优先使用固定模板和结果驱动扩展,不要完全依赖无限制的 AI 猜词。
  • 实质性事实结论必须引用来源:使用 research_pack.md 里的 source ID、URL 或域名。
  • 搜索计划要覆盖多角度:normal 模式下也应尽早包含官方来源和平台特定查询。
  • 搜索预算分阶段使用:先跑种子查询,再跑覆盖查询,最后做结果驱动扩展。
  • 当必要意图已经覆盖,且新增查询的唯一来源产出很低时停止,除非用户明确要求穷尽搜索。
  • 默认不要使用百度视频搜索、图片搜索、秒懂百科视频、阿拉丁或智能搜索生成。本 Skill 是来源检索工具,不是生成答案包装器。

Agent 接入

给其他 Agent 使用时,最稳的默认命令是:

python "{baseDir}/scripts/search.py" "\x3Ctopic>" --mode normal

简单实体定义使用:

python "{baseDir}/scripts/search.py" "\x3Ctopic>" --mode lookup

先读 run_summary.json。百科查询读 lookup_pack.md;网页搜索读 research_pack.md 里的覆盖判断,需要核查原始证据时再读 raw_results.jsonl

Usage Guidance
Install only if you are comfortable giving the skill a Baidu AI Search/AppBuilder API key and sending search terms to Baidu. Use a dedicated key, monitor quota, use --no-cache for sensitive or current searches, and inspect/clear the local runs and cache folders as needed.
Capability Analysis
Type: OpenClaw Skill Name: baidu-api-search Version: 1.0.3 The `baidu-api-search` skill is a legitimate and well-structured tool designed to provide AI agents with advanced Chinese search capabilities using official Baidu APIs (Qianfan and AppBuilder). The implementation uses only Python standard libraries and follows security best practices, such as sanitizing user-provided topics to prevent path traversal (via `slugify_topic` in `adaptive_search.py`) and providing explicit instructions in `SKILL.md` to prevent the AI agent from leaking API keys. The code includes robust features like query planning, result deduplication, and local caching, all of which align with its stated purpose without any signs of malicious intent or data exfiltration.
Capability Tags
cryptocan-make-purchasesrequires-sensitive-credentials
Capability Assessment
Purpose & Capability
The stated purpose—Chinese Baidu/Baike source retrieval, deduplication, and research-pack generation—matches the visible scripts and documentation. The sensitive parts are expected for this purpose: API calls, an API key, and local result files.
Instruction Scope
The skill tells agents to prefer it for Chinese search, recent/latest/hot topics, and Baidu Baike lookup. This is scoped to the skill’s purpose and includes budget guidance such as not using deep mode by default.
Install Mechanism
There is no install spec or remote installer; the visible scripts use Python standard-library style execution. Review confidence is medium because one shown file was truncated and four manifest files were omitted from the supplied file contents.
Credentials
Requiring BAIDU_AI_SEARCH_API_KEYS and making Baidu API requests is proportionate to a Baidu API search skill. Users should still treat search topics as data sent to Baidu and monitor API quota/cost.
Persistence & Privilege
The skill writes run outputs and caches under local run/cache directories. No background service, privileged install, protected-path writes, or self-persistence behavior is shown in the visible artifacts.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install baidu-api-search
  3. After installation, invoke the skill by name or use /baidu-api-search
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.3
Localize the ClawHub-facing skill description, README, and agent prompt to Chinese while keeping a separate English README.
v1.0.2
Add bilingual Chinese and English documentation inside the published skill package.
v1.0.1
Rename to Baidu API Search and align ClawHub slug with API-based positioning.
v1.0.0
Initial release
Metadata
Slug baidu-api-search
Version 1.0.3
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 4
Frequently Asked Questions

What is Baidu API Search?

面向 Agent 的百度 API 搜索 Skill。用于中文“查一下/搜索/联网查/最近/最新/热点”等场景,调用百度百科和百度 AI Search API,保留原始结果、去重并生成 research pack。 It is an AI Agent Skill for Claude Code / OpenClaw, with 132 downloads so far.

How do I install Baidu API Search?

Run "/install baidu-api-search" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Baidu API Search free?

Yes, Baidu API Search is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Baidu API Search support?

Baidu API Search is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Baidu API Search?

It is built and maintained by Valen (@valenovo); the current version is v1.0.3.

💬 Comments