write-mrd

输入一个产品 idea（一句话或详细描述均可），自动生成图文并茂、结构完整的 MRD（市场需求文档），以 Markdown 格式保存到当前项目 markdown/ 目录，同时输出配套 SVG 图表文件。触发条件：用户提到"写 MRD"、"帮我写市场需求文档"、"生成 MRD"、"写产品市场需求"、"我有个产品想法"、"帮我把这个 idea 写成 MRD"、"市场需求文档模板"、"产品立项文档"、"帮我写需求文档"。即使用户只给出一句简短的产品描述如"我想做一个 XX 产品"，也应立即使用本 skill 生成 MRD。本 skill 面向产品经理、创业者、产品负责人，输出可直接用于团队评审和立项决策的专业文档。

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

77%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.

The body is a well-sequenced, highly actionable generation spec with explicit validation (pre-output checklist) and good fallback rules. Its main weaknesses are progressive disclosure — a ~320-line monolith with no reference-file split for the MRD templates and SVG spec — and mild duplication (trigger examples restated in the body, style rules appearing twice).

Suggestions

Split the Step 3 MRD section templates and the Step 4 SVG spec into reference files (e.g. references/mrd-template.md, references/svg-spec.md), keeping SKILL.md as a concise workflow overview with clearly signaled one-level links.

Remove the '触发示例' section — it duplicates the frontmatter description's trigger list and adds no instructional value to the body.

Consolidate the duplicated style rules: the '写作风格规范' section and the '文档质量自检清单' both state the same requirements (e.g. quantification, 'Won't Have', assumption transparency); keep each rule in one place only.

Dimension	Reasoning	Score
Conciseness	Mostly dense and earning its tokens — templates, checklists, naming rules, and hex-color specs rather than explanations of concepts Claude already knows (JTBD/TAM/SAM/SOM/RICE are used as output-structure requirements, not tutorials). Minor trimmable fat remains: the '触发示例' section duplicates the frontmatter trigger list, and style rules appear in both '写作风格规范' and the self-check list. Clearly above anchor 3 (which expects noticeable unnecessary explanation) but not anchor 5 (every token earns its place) given that duplication.	4 / 5
Actionability	Highly actionable instruction-only guidance: copy-paste-ready templates (the 用户角色卡 block, the milestone table, the 文档头部 markdown block), exact file-naming rules ('[product-name]-mrd.md'), exact SVG specs (viewBox="0 0 800 500", font stack, specific hex colors). Anchor 5 is missed by minor gaps — no sample SVG markup despite a detailed SVG spec, and several '根据产品特点选择最关键的维度' judgment calls; comfortably above anchor 3, which would mean pseudocode or missing key details.	4 / 5
Workflow Clarity	A clear 5-step sequence (解析 idea → web 调研 → 构建章节 → SVG → 输出) with an explicit validation checkpoint (the 10-item '文档质量自检清单' to run '在输出前对照检查'), a feedback decision rule in Step 1 (ask one clarifying question only if the idea is under 10 chars with no context, otherwise proceed and mark assumptions), and fallback handling in Step 2 ('如无网络条件，基于已知知识推理，但需…标注数据待验证'). Matches anchor 5: explicit validation steps, error-recovery paths, and a checklist.	5 / 5
Progressive Disclosure	No references/, scripts/, or assets/ exist — everything (~320 lines) is inlined in SKILL.md. Internal structure is good (numbered sections, clear headers), but content that plausibly belongs in reference files (the full MRD section templates of Step 3 and the SVG spec of Step 4) is inline, and there are no well-signaled one-level references. This matches anchor 3 ('content that should be separate is inline') better than anchor 4, whose example assumes the bulk lives in a separate file; above anchor 2 because structure is thorough, not minimal.	3 / 5
	Total	16 / 20 Passed

Description

95%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.

A strong description: concrete capabilities with output location and format, an explicit and comprehensive trigger list in natural user phrasing, and a well-differentiated niche. Its only weaknesses are verbosity — the audience statement and the '即使用户只给出一句简短描述' immediacy instruction add length without adding trigger value — and omission of the research step in the capability summary.

Dimension	Reasoning	Score
Specificity	Names several concrete actions — '自动生成图文并茂、结构完整的 MRD（市场需求文档）', '以 Markdown 格式保存到当前项目 markdown/ 目录', '同时输出配套 SVG 图表文件' — with output format and location. Falls short of anchor 5 because the web-research step is unmentioned and the trailing audience/immediacy statements ('本 skill 面向产品经理、创业者、产品负责人', '即使用户只给出一句简短的产品描述…也应立即使用本 skill') are padding rather than capability statements; clearly above anchor 3, which expects only 1-2 concrete actions.	4 / 5
Completeness	Explicitly answers both what ('自动生成…MRD…保存到当前项目 markdown/ 目录，同时输出配套 SVG 图表文件') and when (a dedicated 触发条件 list of concrete trigger phrases), matching anchor 5 exactly. Anchor 4 ('when could be more explicit') does not apply — the when-clause is as explicit as the good_overall_examples.	5 / 5
Trigger Term Quality	The 触发条件 list covers comprehensive natural synonyms a user would actually say: '写 MRD', '帮我写市场需求文档', '生成 MRD', '写产品市场需求', '我有个产品想法', '帮我把这个 idea 写成 MRD', '市场需求文档模板', '产品立项文档', '帮我写需求文档' — matching the anchor-5 pattern of full synonym coverage including the abbreviated 'MRD' and the full '市场需求文档' forms. No common variation is missing.	5 / 5
Distinctiveness Conflict Risk	Clear niche (MRD / 市场需求文档 generation) with distinct, domain-specific triggers ('写 MRD', '产品立项文档') that would not fire for generic writing or document skills — minimal conflict risk, matching anchor 5. Not below: anchor 4 requires 'minor overlap risk with closely related skills', and no closely related skill shares these triggers.	5 / 5
	Total	19 / 20 Passed

Validation

93%

Warnings & errors only

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

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
relative_links	Relative link issues: 1 missing	Warning

	Total	15 / 16 Passed

Repository: digoal/blog
Path: skills/skills_for_claude_web/write-mrd/SKILL.md
Commit: 69fb793

Reviewed: 8 days ago

Table of Contents

Discovery Implementation Validation

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.