visual-verification

Verify application UI changes with uploaded screenshot or video artifacts

Quality

71%

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

Fix and improve this skill with Tessl

tessl review fix ./packages/sandbox-runtime/src/sandbox_runtime/skills/visual-verification/SKILL.md

Quality

Content

81%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 highly actionable and workflow-clear, with executable command recipes, a success checklist, guardrails, and a cleanup feedback loop. Its main weakness is repetition across the example command blocks, which hurts conciseness and leaves progressive disclosure one step short of ideal.

Suggestions

Factor the shared open/set-viewport/wait/screenshot/upload lines of the three screenshot blocks into one parameterized example to reduce repetition.

Consider moving the long video recording recipe into a references/ file and linking to it from SKILL.md to improve progressive disclosure.

Keep the lean prose style but trim any duplicated flag commentary between 'Recommended Commands' and 'Guardrails'.

Dimension	Reasoning	Score
Conciseness	Prose is lean and assumes Claude's competence, but the four 'Recommended Commands' bash blocks (viewport, full-page, annotated) share roughly 80% identical lines and could be factored into a shared snippet, so it is mostly efficient with tightening possible rather than fully lean.	3 / 5
Actionability	Multiple complete, copy-paste-ready bash blocks with concrete flags, real paths, ffprobe+node probing, and the upload-from-existing-file path cover all common cases, matching the 'Fully executable; specific examples cover the common cases' anchor.	5 / 5
Workflow Clarity	A numbered 'Required Workflow', a 6-item 'Success Criteria' checklist, 'Default Decision Rules', 'Guardrails' acting as validation checkpoints, and a trap-based record-cleanup feedback loop match 'clear sequence with explicit validation steps, feedback loops, and checklists'.	5 / 5
Progressive Disclosure	Content is well-sectioned with clear headers and a single one-level reference to the upload-screenshot skill (no nesting), but at ~200 lines it inlines substantial recipe blocks rather than splitting them into reference files, so it does not reach the ideal 5.	4 / 5
	Total	17 / 20 Passed

Description

61%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 clearly states what the skill does and is reasonably specific and distinct, but it omits an explicit 'Use when...' trigger clause, which caps its completeness and limits trigger-term quality.

Suggestions

Add an explicit trigger clause, e.g. 'Use when verifying a UI change and returning a screenshot or video artifact the user can review.'

Include natural synonyms users say, such as 'before/after', 'visual diff', or 'responsive layout', to broaden trigger coverage.

Name one or two more concrete actions (e.g. 'capture before/after screenshots', 'record interaction flows') to lift specificity toward 4-5.

Dimension	Reasoning	Score
Specificity	Quotes 'Verify application UI changes' and 'uploaded screenshot or video artifacts' name the domain plus two concrete actions/artifacts, but do not enumerate a broader set of specific actions, matching the 'Names domain and 1-2 concrete actions' anchor rather than the several-actions anchor at 4.	3 / 5
Completeness	A clear 'what' is present ('Verify application UI changes with uploaded screenshot or video artifacts') but there is no 'Use when...' clause or equivalent trigger guidance, so per the rubric completeness is capped at 3.	3 / 5
Trigger Term Quality	Terms 'UI changes', 'screenshot', 'video', and 'artifacts' are natural phrases a user would say, giving good coverage; it falls short of 5 because common synonyms like 'before/after', 'visual', 'responsive', or file extensions are absent.	4 / 5
Distinctiveness Conflict Risk	The visual-verification niche with uploaded artifact output is mostly distinct with minimal conflict risk; it sits at 4 rather than 5 because of some overlap with generic screenshot/upload skills.	4 / 5
	Total	14 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: ColeMurray/background-agents
Path: packages/sandbox-runtime/src/sandbox_runtime/skills/visual-verification/SKILL.md
Commit: 52af550

Reviewed: 3 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.