CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

visual-verdict

Structured visual QA verdict for screenshot-to-reference comparisons

Quality

77%

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 ./skills/visual-verdict/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

92%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 an exemplary lean instruction skill: exact output contract, explicit threshold-driven feedback loop, and a worked example with no padding. The only meaningful gap is the absence of scoring criteria for the 0-100 score and guidance for multiple references.

Suggestions

Add brief scoring criteria that map difference severity (layout broken vs. spacing off vs. pixel-level detail) to score bands, so the 0-100 score is deterministic rather than subjective.

Specify how to handle multiple reference_images[] (e.g., judge against the closest-matching reference, or report per-reference differences).

Unify the self-reference naming ('$visual-verdict' vs '/oh-my-claudecode:visual-verdict') to avoid confusion about which invocation is meant.

DimensionReasoningScore

Conciseness

The body is lean with zero padding: every section (output contract, rules, threshold loop, debug aid, worked example) adds information Claude does not already know. Nothing explains general concepts and nothing could be trimmed without losing value.

5 / 5

Actionability

Fully concrete guidance: an exact JSON shape, field-by-field rules, an explicit 90+ pass threshold, and a complete worked example. Minor gaps: no criteria for mapping observed visual differences to the 0-100 score, and handling of multiple reference_images[] is unspecified, which weakens the determinism the skill claims.

4 / 5

Workflow Clarity

The iterate-until-pass loop is explicit with a validation checkpoint: return verdict, check score >= 90, continue editing and rerun, and do not treat the task as complete until the threshold clears. This matches the feedback-loop anchor; the skill is simple and non-destructive, so no further checkpoints are needed.

5 / 5

Progressive Disclosure

A short, single-file skill with clearly labeled sections and no external bundle files; all content is appropriately inline at this size and nothing that belongs in a separate file is present.

5 / 5

Total

19

/

20

Passed

Description

53%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 names a specific niche and a concrete deliverable, but it is a single noun phrase with no trigger clause and no 'Use when' guidance. Adding an explicit when-to-use sentence with natural trigger terms would substantially improve it.

Suggestions

Rewrite with third-person action verbs, e.g., 'Compares generated UI screenshots against reference images and returns a strict JSON verdict (score, differences, suggestions).'

Append an explicit trigger clause: 'Use when the user asks to compare a screenshot/mockup against a reference design, run visual QA, or decide whether a UI matches a target.'

Add natural synonyms users would say (UI, design, mockup, visual diff) to improve trigger-term coverage and distinctiveness.

DimensionReasoningScore

Specificity

Names the domain clearly ('visual QA', 'screenshot-to-reference comparisons') and one concrete deliverable ('Structured... verdict'), but uses no action verbs and covers only a single capability. Fits 'names domain and 1-2 concrete actions' rather than anchor 2, since the deliverable is concrete rather than generic.

3 / 5

Completeness

The 'what' is clear (structured visual QA verdict), but there is no 'Use when...' clause or equivalent explicit trigger guidance — the 'when' is only weakly implied by 'screenshot-to-reference comparisons'. Per the judging guidelines, a missing Use-when clause caps completeness at 3.

3 / 5

Trigger Term Quality

Includes relevant keywords a user might say ('screenshot', 'reference', 'visual QA', 'comparisons') but misses common variations like 'UI', 'design', 'mockup', or 'visual diff'. Some relevant keywords present, but synonym coverage is incomplete.

3 / 5

Distinctiveness Conflict Risk

A clear niche (screenshot-to-reference QA verdicting) that is mostly distinct, with only minor overlap risk against general design-review or visual-comparison skills. Not 5 because triggers are not fully unique to this niche.

4 / 5

Total

13

/

20

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.

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

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

Warning

Total

15

/

16

Passed

Repository

Yeachan-Heo/oh-my-claudecode

Path

skills/visual-verdict/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.