triage-visual-changes

Triage a UI Verify build from your coding agent via the UI Verify MCP — bucket a build's changed stories into real regressions vs cosmetic reflow vs rendering noise, summarize the real regressions for a PR comment, and accept baselines in bulk. Use after a UI Verify build reports changes and you want the agent to review, summarize, or accept from the terminal instead of clicking through the dashboard. Triggers - "triage this build", "summarize the visual regressions", "accept all the new baselines".

Quality

84%

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

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.

A well-structured, actionable skill body with clear multi-step workflows, explicit batch-operation validation, and a useful guardrails checklist. Main gaps are minor redundancy in the before/after guidance and some abstract tool-selector placeholders.

Suggestions

Consolidate the 'render both baseline and candidate, not the after alone' guidance into one canonical statement (e.g., the Guardrails) and reference it from the recipes instead of restating it three times.

Replace abstract tool-selector placeholders ('same selector', 'one of commitSha / prNumber / buildId') with one concrete literal example per tool so invocations are fully copy-paste ready.

Consider extracting the tool table into a references file (e.g. references/tools.md) and linking to it, keeping the recipes as the SKILL.md overview, to improve progressive disclosure for the longer single-file body.

Dimension	Reasoning	Score
Conciseness	Lean and assumes Claude's competence (no explanation of MCP, diffs, or visual regression), but the 'render both before/after, not the after alone' guidance is repeated in Recipe 1, Recipe 4, and Guardrails, which is trimmable redundancy.	4 / 5
Actionability	Provides copy-paste-ready commands ('claude mcp add ...', the Cursor JSON block, 'accept_build { prNumber: 123 }', 'curl -L "$url" -o before.png'), but several tool invocations use abstract selectors ('same selector', 'one of commitSha / prNumber / buildId') rather than literal args, a minor gap.	4 / 5
Workflow Clarity	Each recipe is a clearly sequenced workflow with explicit validation checkpoints for the batch/destructive accept ('confirmed no regression is hiding in the list', 'Read before you write') and a Guardrails checklist with a decision feedback loop (judge flags regression -> adjudicate against before image -> refute or trust), avoiding the batch-cap at 3.	5 / 5
Progressive Disclosure	No bundle files exist, so there are no external references to evaluate; the single self-contained ~107-line file is well-sectioned into Connect MCP, Recipes 1-4, Passed build, and Guardrails with easy navigation, though the embedded tool table is sizeable for an inline block.	4 / 5
	Total	17 / 20 Passed

Description

87%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, specific description with explicit what/when structure and natural trigger phrases tied to a distinct UI-Verify niche. The only real weakness is the second-person voice, which costs it one point on specificity.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions ('bucket a build's changed stories into real regressions vs cosmetic reflow vs rendering noise', 'summarize the real regressions for a PR comment', 'accept baselines in bulk'), which is comprehensive coverage at a 5, but the second-person voice ('your coding agent', 'you want the agent to review') triggers the -1 specificity penalty.	4 / 5
Completeness	Explicitly answers both what (bucket/summarize/accept) and when ('Use after a UI Verify build reports changes and you want the agent to review, summarize, or accept from the terminal') plus concrete trigger phrases.	5 / 5
Trigger Term Quality	Three natural trigger phrases ('triage this build', 'summarize the visual regressions', 'accept all the new baselines') a user would actually say, but no synonyms or file extensions, so it falls short of the comprehensive 5.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear niche ('UI Verify build', 'UI Verify MCP', visual-regression triage) with distinct triggers and minimal overlap with other skills.	5 / 5
	Total	18 / 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: FranciscoMoretti/chat-js
Path: .agents/skills/triage-visual-changes/SKILL.md
Commit: 4584f09

Reviewed: 1 day 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.