CtrlK
BlogDocsLog inGet started
Tessl Logo

dws

管理钉钉产品能力(Agoal目标管理/AI表格/AI搜问/日历/通讯录/群聊与机器人/待办/审批/法务智能合同/考勤/日志/DING消息/开放平台文档/钉钉文档/钉钉云盘/原生Markdown文件/AI听记/邮箱/在线电子表格/知识库等)。当用户需要查询战略解码、经营合约字段、目标规则周期、个人目标、目标模板或周月报规则的按时/迟交/未提交统计与跟催,操作表格数据、管理日程会议、模糊找人/查谁负责某事项、查询通讯录、管理群聊、机器人发消息、创建待办、提交审批、查询或创建合同台账、批量导入合同、按听记起草合同、归档合同、管理合同项目/相对方/账款、查看考勤、提交日报周报(钉钉日志模版)、读写钉钉文档、上传下载云盘文件、读取或修改原生.md文件、查询听记纪要、收发邮件、读写在线电子表格(axls)、管理钉钉知识库,或订阅个人 IM、OA 审批、VoIP 通话邀请、待办或互动卡片回调事件、实时监听群成员加入、群成员退出、群改名和群解散、审批实例发起/抄送/终止/完成、审批任务创建/完成/转交、VoIP 通话邀请、待办创建/更新/删除,以及互动卡片回调时使用。周报/月报正文填报属于日志 Report;规则级提交统计、迟交、未提交和跟催属于 Agoal。

69

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Failed to scan

The risk profile of this skill

SKILL.md
Quality
Evals
Security

Quality

Content

81%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-structured router skill with explicit multi-step workflows, validation checkpoints and a disciplined on-demand reference system. Its main cost is repetition: the schema/help/confirmation contract rules are restated several times across sections, making the 40KB body noticeably heavier than the same guidance needs to be.

Suggestions

Consolidate the confirmation→--yes rule into the 危险操作确认 section and reference it from 严格要求, Shortcut 原则 and Schema 字段速查 instead of restating it in each place; do the same for the leaf-help vs parent-help and 'never inject --all/full leaf into context' rules, which each appear 3+ times.

Move the Schema 字段速查 JSON block and the 'Helper-only 与本地 Cobra 命令' section into references/global-reference.md (or a dedicated schema-reference file), leaving the body with just the progressive-query ladder and the boundary table.

Break the run-on policy paragraphs (e.g., the Shortcut 总览 preamble and the 命令发现 intro) into short bulled rules so each directive is stated once and is scannable.

DimensionReasoningScore

Conciseness

The body is dense and operational with no explanations of concepts Claude already knows, but core directives are repeated verbatim 3-4 times across sections (confirmation→--yes appears in 严格要求, Shortcut 原则, 危险操作确认 and Schema 字段速查; leaf-help precedence and 'don't inject --all/full leaf into context' are each restated in 3+ places). It is not 2 because nothing is conceptual padding, and not 4 because the duplicated policy passages in a 40KB body are a clear tightening opportunity.

3 / 5

Actionability

Commands are copy-paste ready throughout: 'dws schema "calendar event create" --compact', 'dws event consume user_card_action_triggered --flatten -f ndjson', exact error fields ('error.reason == "confirmation_required"', 'retryable/retry_after_seconds/next_retry_at') and a concrete leaf Schema JSON example. It is not 4 because specific executable examples cover the common cases (routing, schema discovery, confirmation retry, event listening) with no material gaps.

5 / 5

Workflow Clarity

A clearly sequenced 4-step core flow (URL precheck → intent classification → ambiguity clarification → product mapping → minimal-read), a 3-step confirmation gate with a full non-interactive retry protocol, and whiteboard flows with mandatory 'whiteboard render' preview / 'whiteboard +diff' validation checkpoints and explicit failure handling. Destructive and batch operations all carry validation (confirm → --yes, --dry-run preview, sourceDigest-guarded update), so the validation cap does not apply; it is not 4 because feedback loops (error → --verbose → report; diff blocked → stop) are explicit rather than implicit.

5 / 5

Progressive Disclosure

A dedicated '详细参考 (按需读取)' section, per-product reference table, documented scripts folder and an explicit minimize-reading policy give good structure, and spot-checks confirm referenced files (references/products/*.md, best_practices/ recipes, scripts/) exist. It is not 5 because some chains are two levels deep (SKILL.md → aitable.md → products/aitable/ with 28 sub-chapters, and dingtalk-tag-index → sub-pages) and some inline material (Schema 字段速查, Helper-only section) reads like global-reference content; it is not 3 because references are clearly signaled and the split is deliberate and navigable.

4 / 5

Total

17

/

20

Passed

Description

96%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

An exceptionally specific and complete description with explicit what/when structure, concrete trigger phrases, extension synonyms, and even cross-product disambiguation (Report vs Agoal). The only weakness is that it is a single ~500-character run-on sentence whose umbrella scope over two dozen products creates overlap risk with the more focused sibling skills it routes to.

DimensionReasoningScore

Specificity

The description enumerates dozens of concrete, verbatim operations such as "批量导入合同", "提交日报周报(钉钉日志模版)", "读写在线电子表格(axls)" and "归档合同" with no vague filler, matching the 'comprehensive coverage' anchor. It is not 4 because the action coverage is exhaustive rather than having minor gaps.

5 / 5

Completeness

It explicitly answers both 'what' ("管理钉钉产品能力……" with enumerated operations) and 'when' ("当用户需要……以及互动卡片回调时使用"), plus concrete trigger phrases and a Report-vs-Agoal disambiguation clause. It is not 4 because the 'when' clause is fully explicit with enumerated concrete triggers, not merely present.

5 / 5

Trigger Term Quality

Natural user phrasings are comprehensively covered ("模糊找人/查谁负责某事项", "周报", "查看考勤", "收发邮件") alongside file extensions and synonyms (".md", "axls", "钉钉文档"), matching the top anchor. It is not 4 because even common variations like 请假/报销-style triggers are addressed by the approval and report phrasings present.

5 / 5

Distinctiveness Conflict Risk

Triggers are clearly DingTalk-specific (product names like 法务智能合同, DING消息, AI听记) giving a clear niche, but the umbrella breadth over 25+ products overlaps with the sibling dingtalk-aitable/dingtalk-chat multi-skills referenced in the body. It is not 5 because that breadth creates real overlap risk with closely related skills; it is not 3 because nothing in the description is generic enough to fire for non-DingTalk skills.

4 / 5

Total

19

/

20

Passed

Validation

68%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 11 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

relative_links

Relative link issues: 48 deeper-than-1-level

Warning

referenced_paths_exist

Referenced path issues: 4 deeper-than-1-level

Warning

Total

11

/

16

Passed

Repository
DingTalk-Real-AI/dingtalk-workspace-cli
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.