CtrlK
BlogDocsLog inGet started
Tessl Logo

write-prd

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 PRD(产品需求文档),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 PRD"、"帮我写需求文档"、"生成 PRD"、"写产品需求文档"、"写商业需求文档"、"我有个产品想法"、"帮我把这个 idea 写成文档"、"PRD 模板"、"产品立项文档"、"产品需求"、"功能需求文档"。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 PRD。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,深度整合 JTBD 用户洞察框架、MoSCoW 优先级方法论和北极星指标体系。

65

Quality

79%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/skills_for_claude_web/write-prd/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

67%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 exceptionally actionable — exact file names, search queries, SVG specs, and a complete PRD template — with a well-sequenced 4-step workflow, scenario handling, and a validation checklist. Its weaknesses are token efficiency (philosophical framing plus a checklist that duplicates template constraints) and the total absence of progressive disclosure: everything, including a 260-line template, is inlined in one 500-line file instead of being split into reference files.

Suggestions

Split the full PRD template (Step 4) and the special-scenario handling (场景 A–F) into files under references/ (e.g., references/prd-template.md, references/scenarios.md), keeping SKILL.md as a lean overview that links to them — this directly addresses the lowest-scoring dimension, progressive_disclosure.

Trim the 写作质量核查清单 to only the checks not already enforced by the template structure (most MoSCoW/JTBD/table checks duplicate what the template mandates), cutting a large block of near-redundant tokens.

Drop the conceptual framing Claude already knows (the WHY→WHAT→HOW→MEASURE philosophy quote and the JTBD worked format), keeping only the instruction to apply the JTBD template when filling section 2.2.

DimensionReasoningScore

Conciseness

The 260-line PRD template and SVG specs are the actual deliverable and earn their tokens, but there is padding a competent model doesn't need: the philosophy framing ('PRD 是团队在不确定性中的共同契约'), a worked JTBD format explainer, and a 30-item 写作质量核查清单 that largely re-states constraints already embedded in the template. Mostly efficient with unnecessary explanation that could be tightened — the anchor-3 fit.

3 / 5

Actionability

Guidance is copy-paste ready throughout: ready-made search query templates ('[产品领域] market size 2024/2025'), exact file-naming conventions ('[product-name]-prd.md', '[product-name]-architecture.svg'), a precise SVG technical spec (viewBox, font-family, full hex color palette), and a complete fill-in PRD template with populated table structures. Nothing is left as vague direction.

5 / 5

Workflow Clarity

A clear 4-step sequence is presented with an ASCII flowchart, six special-scenario branches (A–F), an expected output file manifest, and a post-generation 写作质量核查清单 that functions as an explicit validation checkpoint. It falls short of the anchor-5 'feedback loops for error recovery' pattern only because there is no fix-and-recheck loop after checklist failures.

4 / 5

Progressive Disclosure

The skill is a monolithic ~500-line SKILL.md with no bundle files (no references/, scripts/, or assets/ exist). The full PRD template (~260 lines) and the special-scenario guidance clearly belong in separate reference files that could be loaded on demand, matching the anchor-2 'content that clearly belongs in separate files is inlined'. It is not a 3 because that anchor requires references to be present, and there are none; internal sectioning is decent, keeping it above the worst case.

2 / 5

Total

14

/

20

Passed

Description

91%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: it explicitly states what the skill produces (illustrated PRD in Markdown plus SVG charts, saved to markdown/), when to trigger it, with an exhaustive set of natural Chinese trigger phrases and even the one-sentence-idea edge case. The only weaknesses are minor: a couple of overly broad triggers and the omission of the internal steps (web search, JTBD inference) from the capability list.

DimensionReasoningScore

Specificity

The description names several concrete actions — '自动生成图文并茂、结构完整的 PRD…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' — plus the methodologies used (JTBD, MoSCoW, 北极星指标) and the target audience. It lists multiple specific actions with only minor coverage gaps (e.g., no mention of the web-search step), sitting between the 'several specific actions' (4) and 'comprehensive' (5) anchors, so 4 fits best.

4 / 5

Completeness

Both what ('自动生成图文并茂、结构完整的 PRD…以 Markdown 格式保存…同时输出配套 SVG 图表文件') and when ('触发条件:用户提到…' and '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 clearly and explicitly stated with concrete trigger phrases, matching the top anchor.

5 / 5

Trigger Term Quality

It provides comprehensive natural-language triggers users would actually say — '写 PRD', '帮我写需求文档', '生成 PRD', '写商业需求文档', '我有个产品想法', 'PRD 模板', '产品立项文档', '功能需求文档' — covering synonyms, casual and formal phrasings, plus the edge case '我想做一个 XX 产品'. This matches the 'comprehensive coverage of natural terms including synonyms' anchor exactly.

5 / 5

Distinctiveness Conflict Risk

It carves a clear niche (PRD/product-requirements generation) with distinct trigger phrases, but broader phrasings like '我有个产品想法' or '帮我把这个 idea 写成文档' could overlap with general document-writing or product-brainstorming skills, giving minor overlap risk — matching the 'mostly distinct; minor overlap risk' anchor.

4 / 5

Total

18

/

20

Passed

Validation

87%

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

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (501 lines); consider splitting into references/ and linking

Warning

relative_links

Relative link issues: 3 missing

Warning

Total

14

/

16

Passed

Repository
digoal/blog
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.