CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

visual-artifact-design

创建、编辑、诊断或验证人类可见产物时使用。负责确定自然介质与主要视觉领域、提供轻量工作阶段、选择真实能力、要求当前消费者中的最终像素或等价终态,并让失败返回最早失效步骤;不代替领域 Skill、前端 Skill 或 Runner 的客观运行记录。

60

Quality

76%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./plugins-dist/design-engine/raven_design/skills/visual-artifact-design/SKILL.md
SKILL.md
Quality
Evals
Security

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.

A well-structured process protocol with an explicit state pipeline, validation gates, feedback loops, and exemplary one-level-deep reference bundling. The main costs are token weight from repeated honesty stipulations and long multi-clause policy sentences, plus the absence of a worked example showing a filled-in contract and state record.

Suggestions

Consolidate the honesty/anti-fabrication rules (repeated in §2 states, §6, and the final checklist) into a single short section and reference it, cutting redundant tokens.

Add one compact worked example — a filled-in contract YAML and the corresponding state progression for a typical task — so the abstract states become concretely checkable.

Break up the longest multi-clause sentences (e.g. the identity-inventory and visual-signal paragraphs) into shorter rules; the density currently forces re-reading to extract each individual directive.

DimensionReasoningScore

Conciseness

The body is dense and directive with no filler explaining known concepts, but honesty/anti-fabrication rules recur across §2 ("不得虚报没有发生的参考检查"), §6 ("不得写成独立 PASS"), and the final checklist, and many multi-clause policy sentences (e.g. the identity-inventory paragraph) could be tightened without losing meaning.

3 / 5

Actionability

Concrete, checkable specifics throughout: a YAML contract schema with exact fields, named pipeline states, four named asset routes (truth_asset/standard_symbol/generated_visual/exact_graphic), exact conventions like `https://github.com/<org>.png` and `output_dir=<任务目录>/assets/generated`. Not 5: no filled-in example of the contract or a completed state record, and some criteria (e.g. "三秒记忆点") remain judgment descriptions rather than checkable instructions.

4 / 5

Workflow Clarity

An explicit ordered state pipeline (CONTRACT → … → DELIVERED) with validation gates at each stage (REFERENCE_LOCKED, SURFACES_CLOSED, FINAL_VERIFIED), a dedicated feedback-loop section (§5 RETURN_TO_* with the retention rule "保留仍然有效的母版与检查,不从头重做"), and a final checklist. Error recovery is explicitly sequenced.

5 / 5

Progressive Disclosure

All six referenced files exist in references/, are one level deep (no further nesting inside them), and are each linked from the body — twice contextually (assets-and-imagegen.md, svg-and-vector.md) and once via a clearly signaled on-demand list governed by "仅在会增加消费者合同时读取相关参考". Detailed material (imagegen discipline, SVG rungs, evidence governance) is appropriately split out of SKILL.md.

5 / 5

Total

17

/

20

Passed

Description

67%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.

The description explicitly covers both what and when with a clear boundary disclaimer, but it operates entirely at process-abstraction level. Adding the concrete artifact vocabulary users naturally say (web pages, charts, posters, logos, UI) would sharpen both triggering and distinctiveness.

Suggestions

Add concrete artifact nouns users actually say — e.g. 网页、图表、海报、logo、UI/界面、可视化 — to the when-clause so the skill triggers on natural user phrasing rather than only on the abstract term 人类可见产物.

Name one or two concrete deliverable actions tied to artifacts (e.g. 为网页/海报锁定主视觉母版、生成图标与品牌标识) instead of only process-level actions like 提供轻量工作阶段.

Since the trigger scope (all human-visible artifacts) inherently overlaps domain and frontend skills, state one concrete disambiguation example of when this skill applies versus when a domain skill alone applies.

DimensionReasoningScore

Specificity

Lists several specific actions — "创建、编辑、诊断或验证", "确定自然介质与主要视觉领域", "要求当前消费者中的最终像素或等价终态", "让失败返回最早失效步骤" — but they are process-level rather than artifact-level; no concrete artifact types (web page, chart, poster, logo) are named, so coverage has minor gaps.

4 / 5

Completeness

Both parts are explicit: an explicit when-clause ("…人类可见产物时使用") and a what-clause enumerating responsibilities (stages, capability selection, final-pixel verification, failure return). Not 5 because the when-clause lacks concrete trigger phrases naming artifact types or formats that would make triggering unambiguous.

4 / 5

Trigger Term Quality

Action verbs (create/edit/diagnose/verify) are present, but the core trigger noun "人类可见产物" is a technical abstraction; natural user phrases like "网页", "图表", "海报", "logo", "UI/界面" and file formats are missing. Not 2 because domain-relevant verbs exist, but common variations users actually say are absent.

3 / 5

Distinctiveness Conflict Risk

Explicit boundary statement "不代替领域 Skill、前端 Skill 或 Runner 的客观运行记录" reduces conflict risk, and the scope is a distinct process/governance niche. Not 5 because the trigger scope "人类可见产物" is broad and overlaps closely related frontend, chart, and design skills.

4 / 5

Total

15

/

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

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

Total

14

/

16

Passed

Repository

EverMind-AI/Raven

Path

plugins-dist/design-engine/raven_design/skills/visual-artifact-design/SKILL.md

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.