CtrlK
BlogDocsLog inGet started
Tessl Logo

beautiful-article

把用户提供的素材(网页 URL / PDF / DOCX / Markdown / 纯文本 / 截图 / 粘贴材料)编辑、设计成一篇美丽的、可离线打开和分享的**单文件 HTML 网页文章**。基于 reacticle 组件协议:不手写裸 HTML/CSS,而用语义组件 + 受主题约束的 Raw 自由层;按 source→规划→双确认→生成→终审→修复的小型 harness 流程推进,默认 100% 信息保留的长文。触发场景:把 URL/PDF/DOCX/文章做成网页文章 / 长文 / briefing / 解释文 / 视觉文章 / 教程 / 审阅复盘 / 方案分析,'render this as a beautiful web article / 把这篇做成网页文章 / 生成一篇可分享的 HTML 长文 / reacticle 文章'。只生成文章,不生成后台、表单、dashboard、产品原型或通用 Web App。

72

Quality

88%

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

SKILL.md
Quality
Evals
Security

Quality

Content

77%Weight 40%Scale 1-3

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-engineered, highly actionable multi-phase harness with explicit checkpoints and validation feedback loops. Its main weakness is conciseness (a key decision rule is restated ~5 times) and a progressive-disclosure defect where the repeatedly-cited theme-profiles/ directory is absent from the bundle.

Suggestions

Add the missing theme-profiles/ bundle (theme-profiles/index.json plus per-theme *.md files such as tufte.md and press.md), or remove the theme-profiles references and inline the minimal theme-authoring guidance the body relies on.

Consolidate the '禁止静默替用户选择 / 每项独立确认' rule: state it once in the 质检协议 section and have each Checkpoint reference it by name instead of restating the full rule ~5 times.

Tighten the Phase 3 section — the opening-message template, '如何偏离标配', and '已经走默认值不必单独问' lists overlap; merge the default-value items into the template to reduce token weight.

DimensionReasoningScore

Conciseness

The body is mostly lean and assumes Claude's competence (no explaining of HTML/React/Vite/PDF), but the '禁止静默替用户选择 / 每项独立确认' rule is restated ~5 times (质检协议铁律, Phase 3, Checkpoint 2, Checkpoint 3, 默认策略) and the Phase 3 opening-message template plus '如何偏离标配'/'已走默认值' sections are lengthy. It is efficient with some repetition that could be tightened, which is the level-2 anchor rather than the fully-lean level 3.

2 / 3

Actionability

It gives real executable commands (scaffold.sh with theme/no-cover flags, npm run dev/typecheck/build, html-to-pdf.sh), concrete file paths (article/sections/NN-*.tsx, Article.tsx assembler), and an exact AskQuestion procedure with semantic option labels and a copy-paste opening message. Guidance is concrete and copy-paste ready, matching the top anchor rather than the pseudocode level below.

3 / 3

Workflow Clarity

An explicit Phase 0→8 flow with an ASCII diagram, three hard '必须停' checkpoints, a per-node quality-control table, and feedback loops ('先按 fail 项把产出改完,再汇报…直接拿原始结论汇报但不修复 = 违规'; Section Reviewer pass/fail→fix→report). This is a clear sequence with explicit validation and error-recovery loops — the top anchor — not the level-2 case with missing/implicit checkpoints.

3 / 3

Progressive Disclosure

Organization is strong — a per-phase file-reading guide and a resource table with 'when to read' annotations, one-level-deep references, and a real bundle (references/*.md, article-types/*.md, scripts/*, assets/scaffold-template) matching cited paths. But the body references theme-profiles/<id>.md and theme-profiles/index.json five times as loadable files, and no theme-profiles directory exists in the bundle, so a repeatedly-cited reference path dangles and navigation partially fails. That concrete defect pulls it below the level-3 'easy navigation' anchor into level 2.

2 / 3

Total

10

/

12

Passed

Description

100%Weight 40%Scale 1-3

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

The description is specific, gives explicit trigger terms in both languages, clearly states both what and when, and draws a sharp boundary against adjacent app-building skills. It is a strong, well-scoped description with no significant weaknesses.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — editing/designing material into a single-file HTML article via the reacticle component protocol, running a six-phase harness flow, default 100% info retention — plus an explicit out-of-scope list (no backends, forms, dashboards). It goes beyond naming a domain into comprehensive concrete actions, so it sits at the top anchor rather than the partial-action level below.

3 / 3

Completeness

It states what the skill does (turn material into a shareable single-file HTML article) and provides an explicit trigger clause with literal user phrases, answering both 'what' and 'when'. It is not the level-2 case where 'when' is only implied — triggers are explicit.

3 / 3

Trigger Term Quality

The '触发场景' clause gives broad natural Chinese phrases (长文 / briefing / 解释文 / 视觉文章 / 教程 / 审阅复盘 / 方案分析) plus verbatim user utterances in both languages ('render this as a beautiful web article', '把这篇做成网页文章'). This is good coverage of terms users would actually say, not the sparse keyword level below.

3 / 3

Distinctiveness Conflict Risk

The reacticle-protocol framing and the explicit '只生成文章,不生成后台、表单、dashboard、产品原型或通用 Web App' boundary carve a clear niche unlikely to trigger the wrong skill. It is more sharply distinguished than the 'somewhat specific but could overlap' level below. Voice is third-person/process ('把…编辑、设计成'), so no voice penalty applies.

3 / 3

Total

12

/

12

Passed

Validation

93%

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

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 1 missing

Warning

Total

15

/

16

Passed

Repository
ConardLi/garden-skills
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.