AI搜问:人员语义搜索、跨源主题检索与行为回溯。Use when 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch。

Apache-2.0Auto-check passed

Install Dingtalk Aisearch

skills CLI
$ npx skills add DingTalk-Real-AI/dingtalk-workspace-cli --skill dingtalk-aisearch -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install DingTalk-Real-AI/dingtalk-workspace-cli dingtalk-aisearch --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/multi/dingtalk-aisearch .claude/skills/dingtalk-aisearch && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
dingtalk-aisearch
GitHub stars
3.2k
Token cost
~1.2k tokens
SKILL.md length
260 words
Files
4 (incl. references)
Skills in repo
17
Repo updated
First seen
Licence
Apache-2.0

At a glance

AI搜问:人员语义搜索、跨源主题检索与行为回溯。Use when 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch。

  • Works in 6 steps: 人员搜索 → 跨源内容搜索 → 行为回溯 → …
  • 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch
  • SKILL.md covers 最小执行契约, 选路, 1. 人员搜索 and 2. 跨源内容搜索, plus 6 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Dingtalk Aisearch is an agent skill from DingTalk-Real-AI/dingtalk-workspace-cli. AI搜问:人员语义搜索、跨源主题检索与行为回溯。Use when 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch。

Its SKILL.md is about 1.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 4 other files, including reference files (for example `references/aisearch.md`, `references/intent-guide.md` and `references/lite-recipes.md`).

The repository describes itself as: DingTalk Workspace is an officially open-sourced cross-platform CLI tool from DingTalk. It unifies DingTalk’s full suite of product capabilities into a single package, is… The licence is Apache-2.0.

When your agent uses it

  • 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch

Example prompts

  • “/dingtalk-aisearch”

Workflow steps

6 steps, taken from the step headings in SKILL.md.

  1. 人员搜索
  2. 跨源内容搜索
  3. 行为回溯
  4. 核验与交付
  5. 同条件原生核验与多跳
  6. 禁止本地兜底

What it can do on your machine

Read from SKILL.md and the folder at commit 0a813a9. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md (its code samples are bash).

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Dingtalk Aisearch loads about 1.2k tokens when it runs, and up to ~3.4k if it reads all its reference files. Until then it costs about 32 tokens; SKILL.md has 260 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~32
When it runs · the whole SKILL.md, loaded when a task matches
~1.2k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~3.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from DingTalk-Real-AI/dingtalk-workspace-cli at commit 0a813a9, republished under its Apache-2.0 licence (© DingTalk-Real-AI). 260 words, ~1,242 tokens.

Download SKILL.mdSave it as .claude/skills/dingtalk-aisearch/SKILL.md (or your agent's skills folder). This skill also uses 3 other files; get the full folder from GitHub.
name
dingtalk-aisearch
description
AI搜问:人员语义搜索、跨源主题检索与行为回溯。Use when 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch。
metadata.cli_version
>=0.2.14
metadata.category
product

钉钉 AI 搜问 Skill

<!-- DWS_RUNTIME_CONTRACT_START -->

最小执行契约

  • 只用 dws 操作钉钉,每条命令带 --format json,只按结构化返回下结论。
  • person/enterprise/behavior 直接调用,不预读 shared、Reference、Schema、Help 或下游 Skill。
  • 不猜命令、字段、ID、profile 或事实;多候选不取首项,不混用不同 domain 的 ID。
  • 空结果结束本次搜索;失败、不完整或候选结果不能表述成“对象不存在”、全量或唯一。
  • 来源不确定或用户列出多个来源时,第一条业务命令必须是一次合并 --types 的 aisearch enterprise,不能先逐产品搜索。
<!-- DWS_RUNTIME_CONTRACT_END -->

选路

意图入口
姓名、工号、部门、职位、职责、上下级或手机号线索找人dws aisearch person
按主题跨文档、消息、邮件、待办、日程、听记等找内容dws aisearch enterprise
以我为端点的发送/接收,或我创建、编辑、分享过什么dws aisearch behavior
<!-- dws-intent: chat.search.filtered -->仅 IM 且需逐条消息谓词过滤dws chat +search-msg
无主题地列最近访问/编辑文档dws drive +recent
枚举部门完整成员dingtalk-contact
完整手机号精确反查dws contact user search-mobile --mobile "<手机号>" --format json
已知稳定 ID 后读取/修改对应产品 Skill

顺序:资源范围 → 答案形态 → 原生谓词。

1. 人员搜索

维度:姓名=name,部门=department,职位=position,职责/负责人=duty,上级=supervisor,下属=subordinate,工号=jobNumber,手机号线索=phone;确实无法判断才用 all。

bash
dws aisearch person --query "<用户原始目标>" --dimension <维度> --format json
  • 保留完整目标;独立条件分别查。正确维度为空就结束,不换 all、缩词或扩同音词。
  • 用 searchEvidence.returnedCandidateCount 报告本次候选数;无分页完成证据只称“本次返回”。
  • 姓名/花名包含查询只交付 _searchEvidence.exactAliasContainsQuery=true;exactAliasContainsCount=0 即无精确包含项,语义近似不能替代。
  • resolvedRelation.status=resolved 时直接使用 target/supervisor/relation 回答直属上级,不再搜索,也不把上级的上级当直属上级。
  • 产品归属与负责人用一次 dimension=duty,保留产品名和“负责”;仅职责证据明确者可称负责人,多候选就列候选及部门,共同上级不等于负责人。
  • 姓名与花名可由同一候选的 aliases 关联;不同域 ID 不可直接比较。需要更多人员详情时才用真实 userId 切 Contact。

2. 跨源内容搜索

时间词只进 --time-range,类型词只进 --types,其余主题进 --queries。

bash
dws aisearch enterprise --queries "<主题>" --types <类型CSV> [--time-range "<原时间词>"] --format json

类型:文档/普通文件=document,消息=im,邮件=mail,日程=calendar,待办=todo,听记=minute,日志=report,AI表格=notable,知识内容=baike;用户列出的来源合并到一次调用。

  • requestConstraints 回显实际条件,sources 给出各来源数量;no_returned_items 仅表示本次未命中,coverage.status=unknown 不能证明完整。
  • _searchEvidence.queryMatch=text_evidence_present 才有文本主题证据;no_text_evidence 仅是语义候选,标为未核实。
  • 只交付 result 保留项并说明 delivery.omitted;doNotRetryOrExpand=true 时停止,不改词、逐产品补搜或读 Help。
  • 唯一精确标题可能附带 resolvedDetail。其状态为 resolved 或 resolved_from_search_payload 且字段足够时直接回答;failed 才报告补充读取失败。
  • 日程空命中时接口可按原时间调用日历,meta.nativeFallback=true 仍属本次 AISearch 证据,无需 Agent 再查日历。
  • 精确标题未命中不拿近似标题、最近项或首项替代。“唯一才读取”须有指定来源覆盖、分页结束和唯一精确匹配;正文 complete=true 不代表搜索完整。
  • 用户只要候选摘要或链接时不读原文;确需正文且同次结果缺字段,才按真实稳定 ID 读取对应产品。

3. 行为回溯

bash
dws aisearch behavior --queries "<主题>" --types <类型CSV> --behavior-type <all|send|receive|create|edit|share> [--time-range "<原时间词>"] [--direction "我->某人|某人->我|我<->某人"] [--chat-scope "<完整群名>"] --format json
  • 每组动作、方向、时间调用一次并合并类型。方向按当前用户视角:“我发给某人”=send;“某人发给我”=receive。chat-scope 仅用于 IM,方向保留原姓名。
  • 成功空数组就交付“已查询,本次 0 条”,按来源列 0;不再用 enterprise 拼行为证据,不删主题、扩时间、缩群名或用 recent 代替。
  • timeEvidence.withinRequestedRange=false 排除;为 null 则标时间未核实。messageEvidence.perMessageVerified=false 时不能把混合会话 snippet 全归给目标人。
  • “某人发过 Word 文件”或“某人今天发给我的消息”只调用一次 behavior:types=im、receive、人物方向及原时间词;不切 Chat、猜 sender 或扩大年份。
  • resolvedMessageEvidence.status=resolved 时直接使用 messages,即使为空也不再调用 Chat;identity.match=returned_alias 只证明展示名匹配返回别名。
  • 发送者未知的“谁给我发过某主题文件”用 direction="某人->我",按 identity.match=any_returned_sender 的 messages.sender 汇总,不先猜人或拆两次搜索。

4. 核验与交付

  • 时间:保留原时间词;“本周”不等于近七天。确认数值时间单位。创建行为用创建时间,不能用修改时间或会议开始时间代替。
  • 身份:核对发送者、收件关系和创建者;群内可见不等于发给我。优先用 stableRefs/identityRefs,仅同 domain 的 ID 可直接比较。
  • 主题/类型:须有文本、附件或对象类型证据;群名、普通链接和搜索命中本身不能证明主题或文件属性。
  • 唯一/全量:检查来源和分页终态,按稳定 ID 去重;只问“有没有”时,有效正向命中即可。
  • 摘要:不补写未返回字段,不把建议写成事实,不把 snippet 称正文;分别报告命中、空结果、失败和未核实项。

5. 同条件原生核验与多跳

路径:AISearch 定位 → 确认目标 → 优先使用同次 resolvedDetail → 缺必需字段才按真实 ID 调用一个对应产品。

  • 仅在需要详情、完整成员或证据核验,且已有稳定 ID 或原生可执行的相同精确标题/部门/范围时转用对应产品;保持主题、身份、方向、时间和权限,不扩大扫库。
  • 跨来源第一步固定为 enterprise。无稳定 ID 时不能用原生重搜替代;目标不明、身份矛盾、失败或无权限时说明限制并停止。
  • 人员用 userId,文档用 nodeId,听记用 taskUuid,待办用真实 taskId;Ding uid 不能当 Contact userId,snippet 不能当正文。
  • 用户的“唯一才读取”等前置条件仍须满足。ID 域或下一跳不明确时才读 多跳短流程。

6. 禁止本地兜底

钉钉对象只用 DWS 返回作业务证据;未命中时不得用 find、rg、mdfind 或评测 fixture 回答。会议议题负责人用一次 person --dimension duty 保留会议与职责原词;只有内容能提供职责原文时才改用一次 enterprise --types calendar,im,document,二者不都跑。

错误与成本

  1. retryable=true 时原调用最多重试一次;否则停止,不换身份或绕权限。
  2. 成功但为空或无精确目标:不重复、改词或扩时间。
  3. unknown flag 才查看一次该命令 Help;API、权限和空结果不靠 Help 解决。
  4. 独立来源可继续;依赖失败结果的后续停止。保留来源、链接、ID、范围、数量、完整性和错误。

按需 Reference

常用路径不读 Reference。低频枚举或兼容参数不明才读 完整命令参考;产品边界不明读 局部意图消歧;稳定 ID 域或下一跳不明读 多跳短流程。

© DingTalk-Real-AI, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 3 other files (references) in skills/multi/dingtalk-aisearch of DingTalk-Real-AI/dingtalk-workspace-cli.

  • SKILL.md
  • references/aisearch.md
  • references/intent-guide.md
  • references/lite-recipes.md

Open the folder on GitHubat commit 0a813a9

Compare with similar skills

Dingtalk Aisearch next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Dingtalk Aisearch compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Dingtalk Aisearch this skillDingTalk-Real-AI/dingtalk-workspace-cli3.2k—~1.2kAutomated safety check: PassApache-2.0
Dingtalk Work Logmohitagw15856/pm-claude-skills1.4k—~889Automated safety check: PassMIT
Dingtalk Messageagentscope-ai/ReMe3.6k—~1.6kAutomated safety check: PassApache-2.0
Dingtalk TeambitionLeoYeAI/openclaw-master-skills2.2k—~3.7kAutomated safety check: PassMIT
Dingtalk Channel RulesDingTalk-Real-AI/dingtalk-openclaw-connector2.1k—~595Automated safety check: PassMIT
Dingtalk Bridgeaiskillstore/marketplace433—~1.2kAutomated safety check: NotesMIT

Similar skills

  • Dingtalk Work Log

    mohitagw15856/pm-claude-skills

    Write a DingTalk work log (钉钉日志): the daily report (日报), weekly report (周报) or monthly report (月报) in the fields DingTalk's built-in log templates use, so it can be filled in directly.

    1.4k GitHub stars~889 tokensUpdated yesterday
    Auto-check passed
  • Dingtalk Message

    agentscope-ai/ReMe

    钉钉消息发送技能。支持企业内部机器人(批量单聊/群聊)和 Webhook 自定义机器人两种接入方式,支持多机器人管理,支持文本、Markdown、链接、ActionCard、FeedCard等多种消息类型。

    3.6k GitHub stars~1.6k tokensUpdated yesterday
    Backend & APIsAuto-check passed
  • Dingtalk Teambition

    LeoYeAI/openclaw-master-skills

    A skill your agent uses for anything related to Teambition tasks and projects.

    2.2k GitHub stars~3.7k tokensUpdated 2 mo ago
    DevelopmentAuto-check passed
  • Dingtalk Channel Rules

    DingTalk-Real-AI/dingtalk-openclaw-connector

    钉钉会话输出规则。在钉钉会话中始终生效。

    2.1k GitHub stars~595 tokensUpdated 14 days ago
    Auto-check passed
  • Dingtalk Bridge

    aiskillstore/marketplace

    DingTalk group chat bridge for Claude Code. An agent skill from aiskillstore/marketplace.

    433 GitHub stars~1.2k tokensUpdated today
    Productivity & AutomationAuto-check: notes
  • Dingtalk Unified

    infometa/workbuddyskills

    钉钉 CLI 全能套件,基于官方 DingTalk Workspace CLI(dws)操作钉钉消息、群聊、通讯录、日历、待办、审批、考勤、日志、DING、AI 表格、钉钉文档、钉盘、AI 听记、邮箱和开放平台文档。用户需要在钉钉内查找联系人、发消息、建群、管理日程、创建待办、处理审批、查看考勤、读写文档、操作 AI 表格、搜索会议听记或调用钉钉开放平台能力时使用。

    348 GitHub stars~3.9k tokensUpdated today
    Auto-check passed

More from DingTalk-Real-AI/dingtalk-workspace-cli

All 17 skills in this repo
  • Dingtalk Minutes

    DingTalk-Real-AI/dingtalk-workspace-cli

    钉钉 AI 听记。Use when 查询或修改听记摘要、完整逐字稿、关键词、标签、行动项、录音、上传、思维导图、发言人洞察、ASR 热词/识别词配置或分享权限。写文档走 dingtalk-doc;建待办走 dingtalk-todo;日程走 dingtalk-calendar。命令前缀:dws minutes。

    3.2k GitHub stars~2.3k tokensUpdated today
    Auto-check passed
  • Dingtalk Aitable

    DingTalk-Real-AI/dingtalk-workspace-cli

    钉钉 AI 表格(多维表)业务操作。Use when 用户需要操作 AI 表格/多维表/Base/Table、应用模式/App 页面/Widget、建表、查写记录、字段、记录评论、评论回复、访问密钥(API Key)、SQL/PostgreSQL/SELECT/JOIN、筛选、排序、公式、模板、批量导入 CSV 或 JSON、导出、仪表盘、图表、上传附件、数据源配置与同步、按任务 ID…

    3.2k GitHub stars~5.4k tokensUpdated today
    Auto-check passed
  • Dingtalk Aicard

    DingTalk-Real-AI/dingtalk-workspace-cli

    Create, edit, and validate DingTalk AI Card (A2UI) JSON files offline.

    3.2k GitHub stars~2.5k tokensUpdated today
    Auto-check passed
  • Dingtalk Calendar

    DingTalk-Real-AI/dingtalk-workspace-cli

    钉钉日历与会议室。Use when 用户说 约会议/查日程/订会议室/查闲忙/加参会人/改期/取消会议/今天的日程/本周日程/共同空闲。不做视频会议发起/邀请入会/会中控制(走 dingtalk-misc)、AI 听记(走 dingtalk-minutes)、待办任务(走 dingtalk-todo)。命令前缀:dws calendar。

    3.2k GitHub stars~2k tokensUpdated today
    Auto-check passed
  • Dingtalk Contact

    DingTalk-Real-AI/dingtalk-workspace-cli

    钉钉通讯录精确查询。Use when 已有 userId 后查详情、部门、职位或邮箱,按完整手机号反查用户,或查询自己、部门成员及角色。姓名模糊搜索、工号、职责、上下级走 dingtalk-aisearch,拿到 userId 后用本 skill 补详情。命令前缀:dws contact。

    3.2k GitHub stars~2.2k tokensUpdated today
    Auto-check passed
  • Dingtalk Mail

    DingTalk-Real-AI/dingtalk-workspace-cli

    钉钉邮箱读写、搜索、回复与转发。Use when 用户说发邮件/查邮件/回邮件/转发邮件/未读邮件/邮件搜索/邮箱附件。一句话发邮件时先用 dingtalk-contact 解析并确认收件人,再由本 skill 发送;不做钉钉消息(走 dingtalk-chat)、紧急通知(走 dingtalk-misc)。命令前缀:dws mail。

    3.2k GitHub stars~1.4k tokensUpdated today
    Auto-check passed

Questions about Dingtalk Aisearch

What does Dingtalk Aisearch do?

AI搜问:人员语义搜索、跨源主题检索与行为回溯。Use when 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch。. Dingtalk Aisearch is an agent skill from DingTalk-Real-AI/dingtalk-workspace-cli.

When should I use Dingtalk Aisearch?

Dingtalk Aisearch fits situations like: 语义找人,或目标未知时按主题或行为发现内容。原生最近列表走所属产品;完整手机号精确反查走 dingtalk-contact。前缀:dws aisearch.

How do I install Dingtalk Aisearch in Claude Code?

Run `npx skills add DingTalk-Real-AI/dingtalk-workspace-cli --skill dingtalk-aisearch -a claude-code`. Or copy the skill folder (skills/multi/dingtalk-aisearch in DingTalk-Real-AI/dingtalk-workspace-cli) into .claude/skills/dingtalk-aisearch in your project. Claude Code loads it when a task matches its description.

How do I install Dingtalk Aisearch in Codex?

Run `npx skills add DingTalk-Real-AI/dingtalk-workspace-cli --skill dingtalk-aisearch -a codex`. Or copy the skill folder (skills/multi/dingtalk-aisearch in DingTalk-Real-AI/dingtalk-workspace-cli) into .agents/skills/dingtalk-aisearch in your project. Codex loads it when a task matches its description.

Can I use Dingtalk Aisearch in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add DingTalk-Real-AI/dingtalk-workspace-cli --skill dingtalk-aisearch -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/dingtalk-aisearch, .gemini/skills/dingtalk-aisearch, .github/skills/dingtalk-aisearch and .opencode/skills/dingtalk-aisearch in your project.

What does Dingtalk Aisearch need to run?

SKILL.md names no scripts, command-line tools or credentials: Dingtalk Aisearch is instructions for the agent only.

Does Dingtalk Aisearch access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Dingtalk Aisearch safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Dingtalk Aisearch use?

Dingtalk Aisearch is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Dingtalk Aisearch use?

About 1.2k tokens (SKILL.md is roughly 5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.1k tokens, read only when the agent opens those files.

What are the alternatives to Dingtalk Aisearch?

Skills that share tags, products or a category with Dingtalk Aisearch: Dingtalk Work Log (mohitagw15856/pm-claude-skills, 1.4k stars), Dingtalk Message (agentscope-ai/ReMe, 3.6k stars), Dingtalk Teambition (LeoYeAI/openclaw-master-skills, 2.2k stars) and Dingtalk Channel Rules (DingTalk-Real-AI/dingtalk-openclaw-connector, 2.1k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Dingtalk Aisearch?

DingTalk-Real-AI (a GitHub organization) maintains it in DingTalk-Real-AI/dingtalk-workspace-cli, which has 3,246 GitHub stars. The repository holds 17 skills in this directory. The repository was last updated on October 10, 2026.

Source: DingTalk-Real-AI/dingtalk-workspace-cli on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.