agent-browser-verify

Automated browser verification for dev servers. Triggers when a dev server starts to run a visual gut-check with agent-browser — verifies the page loads, checks for console errors, validates key UI elements, and reports pass/fail before continuing.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

88%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 highly actionable, well-sequenced verification workflow with executable commands, explicit validation checkpoints, and a bounded retry loop. The only notable improvements are deduplicating repeated eval snippets and moving the specialized server-log correlation material into a reference file.

Suggestions

Deduplicate the repeated eval checks (window.__consoleErrors and the error-overlay selector appear in both the Quick Verification Flow and the Diagnosing sections) by referencing one canonical snippet.

Move the server-side correlation section (vercel logs, npx workflow inspect/health, and the correlation table) into a references/ file for stuck-page diagnosis, keeping SKILL.md focused on the core verification flow.

Dimension	Reasoning	Score
Conciseness	The body is efficient — terse command blocks, a compact checklist, and a symptom-to-cause table with no tutorials on concepts Claude already knows. It misses the 5 anchor due to small redundancies: the window.__consoleErrors eval and the error-overlay eval each appear twice, and phrases like "the browser is only half the story" are slight padding.	4 / 5
Actionability	Every step is copy-paste executable: agent-browser commands, concrete selectors ([data-nextjs-dialog], .vite-error-overlay, #webpack-dev-server-client-overlay), runnable eval expressions, and specific fixes in the correlation table (vercel link && vercel env pull, increasing maxDuration). This fully matches the top anchor.	5 / 5
Workflow Clarity	The sequence is explicit (open → wait for networkidle → screenshot → error check → snapshot), backed by a 6-item verification checklist, a numbered failure path, a bounded retry loop ("max 2 retry cycles to avoid infinite loops"), and a re-verify procedure with feedback recovery. Matches the 5 anchor with explicit validation checkpoints.	5 / 5
Progressive Disclosure	The single-file body has clear, well-ordered sections and no dangling references (no bundle files exist). It is not a 5 because the server-side correlation deep-dive (vercel logs, npx workflow inspect/health) is niche troubleshooting that would sit more appropriately in a separate reference file.	4 / 5
	Total	18 / 20 Passed

Description

88%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 description that clearly states concrete capabilities and an explicit trigger condition in third-person voice. The main improvement opportunity is surfacing a few natural user phrases (blank page, screenshot, page not loading) in the description text itself rather than only in metadata prompt signals.

Suggestions

Fold one or two natural user phrasings into the description (e.g., "...or when the user reports a blank page, stuck spinner, or console errors") to strengthen trigger-term coverage.

Name the supported dev server types (Next.js, Vite, etc.) in the description to further sharpen distinctiveness from generic browser-testing skills.

Dimension	Reasoning	Score
Specificity	The description lists multiple concrete actions — "verifies the page loads, checks for console errors, validates key UI elements, and reports pass/fail" — giving comprehensive coverage of the verification workflow. It exceeds the 4 anchor because there are no significant gaps in the action coverage.	5 / 5
Completeness	It explicitly answers both questions: the "what" is the four concrete verification actions with pass/fail reporting, and the "when" is stated directly as "Triggers when a dev server starts". This matches the top anchor rather than the 4 anchor because the trigger is explicit, not merely implied.	5 / 5
Trigger Term Quality	"Triggers when a dev server starts" plus terms like "console errors", "page loads", and "browser" provide good keyword coverage. It falls short of the 5 anchor because common natural phrases users would say ("blank page", "screenshot the page", "page not loading") appear only in the metadata promptSignals, not in the description itself.	4 / 5
Distinctiveness Conflict Risk	The dev-server verification niche with a named tool (agent-browser) is mostly distinct. It is not a 5 because there is minor overlap risk with generic browser-automation, screenshot, or web-testing skills.	4 / 5
	Total	18 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning

	Total	14 / 16 Passed

Repository: openai/plugins
Path: plugins/vercel/skills/agent-browser-verify/SKILL.md
Commit: 5fd93af

Reviewed: 2 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.