← Back to Skills Marketplace
168
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install case-search
Description
根据用户描述的实际场景检索并匹配最相关的案例，返回最多三个符合需求的详细判决信息。
README (SKILL.md)
执行流程
step1: 根据用户场景找到类似匹配的案例并返回案例号
- 如果查找财产类案件请访问~/.openclaw/workspace/skills/case_search/case_info/case_index.md目录文件，并根据目录指引访问该文件夹下案件详情文件
stp2: 匹配概括信息，找到最相近的案件
用"概括"字段匹配用户当前场景和需求，返回最相关的案例(最多3个)，如果没有符合当前场景的类似案例请返回"无"
step2:查询案例详情
step2-1:若有查到符合当前场景的类似案例
根据最相关的案号(case_no)调用sql_helper skills查询数据库中company_db.case_info表中的数据返回案情详情
step2-2:若没有查到符合当前场景的类似案例
请调用tavily_search skills技能进行网络检索查询相关案情
step3:根据查询结果结构化输出案例详情
若找到最相关的案例:
【案号】:XXXX 【法院】:XXXX 【审理时间】:XXXX 【事实与理由】:XXXX 【裁判依据】:XXXX 【本院认为】:XXXX 【判决如下】:XXXX
【案号】:XXXX 【法院】:XXXX 【审理时间】:XXXX 【事实与理由】:XXXX 【裁判依据】:XXXX 【本院认为】:XXXX 【判决如下】:XXXX ...
若无最相关案例
【案号】:无 【法院】:无 【审理时间】:无 【事实与理由】:无 【裁判依据】:无 【本院认为】:无 【判决如下】:无 (请务必严格按上述格式输出，请勿做任何解释说明,请勿输出案例内容)
Usage Guidance
This is an instruction-only skill that bundles local case summaries and asks the agent to match and return up to three relevant cases. Before installing: (1) note the skill will call other skills (sql_helper to query a database and tavily_search to search the web) — confirm those skills are trustworthy since they may require DB credentials or network access; (2) the SKILL.md has minor contradictions (it both asks to return detailed case info and instructs 'do not output case content') — decide which behavior you want and test the skill in a sandbox first; (3) the skill reads files under ~/.openclaw/workspace/skills/case_search/ — ensure your agent uses that workspace layout and that exposing those files is acceptable; (4) if you are concerned about database queries, restrict or review sql_helper's permissions instead of relying on this skill to enforce them.
Capability Analysis
Type: OpenClaw Skill
Name: case-search
Version: 1.0.0
The skill bundle is a legitimate legal case search tool designed to retrieve and format judicial precedents. The instructions in SKILL.md guide the agent to search local markdown files in the case_info/ directory and use supplementary tools like sql_helper (targeting a specific company_db.case_info table) or tavily_search for broader queries. There is no evidence of malicious intent, data exfiltration, or unauthorized execution; the behavior is entirely consistent with the stated purpose of legal research.
Capability Assessment
Purpose & Capability
Name/description match the included case_info files and the declared purpose of retrieving relevant cases. Minor inconsistency: the skill includes local case files and instructs the agent to read them, yet it also directs calling an external sql_helper skill to query company_db.case_info for case details—this is unnecessary given the packaged files and may change the trust boundary. Also the SKILL.md's final instruction says '请勿输出案例内容' (do not output case content) while earlier it requests returning detailed judgment information—this is contradictory.
Instruction Scope
Runtime instructions are explicit and scoped: read the skill's case_info files under the agent workspace and match by the '概括' field; if matches found, call sql_helper to fetch detailed case_info by case_no; if none, call tavily_search for web search. The instructions do not ask to read arbitrary system files or environment variables. The main scope concern is dependence on other skills (sql_helper and tavily_search), which may access databases or the network outside this skill's control.
Install Mechanism
No install spec and no code files to execute; the skill is instruction-only with bundled static markdown case files—this is low-risk from an install/execution perspective.
Credentials
The skill declares no required env vars or credentials (proportionate). However, it instructs calling sql_helper (which will likely require DB credentials) and tavily_search (which performs web queries). Those credentials/permissions would be supplied to the other skills, not to this one—verify sql_helper/tavily_search are trusted and that you accept granting them any DB/network access.
Persistence & Privilege
No high-privilege flags (always:false). The skill uses default autonomous invocation capability, which is normal for skills. It does not request persistent system-wide changes or modify other skills' configurations in the provided instructions.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install case-search - After installation, invoke the skill by name or use
/case-search - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Case skill versiond 1.0.0 – Initial Release
- Introduced structured legal case search and retrieval functionality.
- Added multiple case type modules including corruption/bribery, finance (divorce/eat/steal), human injury, and traffic accident.
- Implemented workflow to identify and return the most relevant legal cases based on user scenarios.
- Provided formatdted, structured outputs for case details or indicated when no relevant case is found.
- Replaced general-purpose documentation with focused instruction for case search usage.
Metadata
Frequently Asked Questions
What is case_search?
根据用户描述的实际场景检索并匹配最相关的案例，返回最多三个符合需求的详细判决信息。 It is an AI Agent Skill for Claude Code / OpenClaw, with 168 downloads so far.
How do I install case_search?
Run "/install case-search" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is case_search free?
Yes, case_search is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does case_search support?
case_search is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created case_search?
It is built and maintained by zhuwano1 (@zhuwano1); the current version is v1.0.0.
More Skills