agent-browser

Browser automation CLI for AI agents. Use when the user needs to inspect, test, or automate browser behavior: navigating pages, filling forms, clicking buttons, taking screenshots, extracting page data, reading selected Open Design browser-tab context, testing web apps, dogfooding Open Design previews, QA, bug hunts, or reviewing app quality. Prefer local Open Design preview URLs unless the user explicitly asks for external browsing.

Quality

94%

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

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.

A highly actionable, well-structured skill body with executable commands, explicit validation checkpoints, and clean deferral of upstream guides to temp files. The only weakness is mild token redundancy from the repeated CDP startup block across two sections.

Suggestions

Factor the shared CDP startup sequence into a single canonical snippet referenced from both 'CDP Startup Contract' and 'Open Design Smoke Path' to remove the near-duplicate block and improve conciseness.

In the numbered Workflow, add an explicit retry/fallback step for failed smoke-path assertions (e.g., expected title/URL mismatch) to make the feedback loop for the smoke path as explicit as the CDP recovery loop.

Dimension	Reasoning	Score
Conciseness	The body is dense and operational with no padding about browser/CDP basics, but the CDP startup block is duplicated nearly verbatim in both 'CDP Startup Contract' and 'Open Design Smoke Path', a minor trim opportunity that fits 'efficient; minor instances that could be trimmed' (level 4) rather than 'every token earns its place' (level 5).	4 / 5
Actionability	Provides fully executable, copy-paste-ready bash (command -v agent-browser, the full CDP startup/poll/connect sequence, agent-browser snapshot/screenshot) plus exact expected-success criteria and exact error strings to report, matching the level-5 'fully executable; specific examples cover the common cases' anchor.	5 / 5
Workflow Clarity	An explicit 11-step Workflow is backed by a self-contained smoke path with validation checkpoints (CDD readiness polling, expected title/URL/snapshot/screenshot checks) and real error-recovery feedback loops (crash -> manual launch -> report DevToolsActivePort), satisfying the level-5 anchor for explicit validation and feedback loops.	5 / 5
Progressive Disclosure	No bundle files exist (references/scripts/assets absent); the body is well-organized into clear sections and appropriately defers specialized detail to one-level-deep, clearly-signaled runtime-fetched guides ('agent-browser skills get ... > temp file', 'Load these only when directly needed'), matching the level-5 anchor for well-signaled one-level-deep references.	5 / 5
	Total	19 / 20 Passed

Description

96%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 that clearly states what the skill does and when to use it, with comprehensive natural trigger terms. The only soft spot is distinctiveness, since the broad browser-automation/QA framing has some overlap risk with adjacent web-testing skills despite the Open Design narrowing.

Dimension	Reasoning	Score
Specificity	Lists many concrete actions ('navigating pages, filling forms, clicking buttons, taking screenshots, extracting page data, reading selected Open Design browser-tab context, testing web apps, dogfooding Open Design previews, QA, bug hunts, or reviewing app quality') with comprehensive coverage across the browser-automation domain, matching the level-5 anchor rather than the 'minor gaps' of level 4.	5 / 5
Completeness	Explicitly answers both 'what' ('Browser automation CLI for AI agents... inspect, test, or automate browser behavior') and 'when' ('Use when the user needs to inspect, test, or automate browser behavior...') with concrete trigger phrases, matching the level-5 anchor.	5 / 5
Trigger Term Quality	Includes natural user phrases ('fill out a form', 'click a button', 'take a screenshot', 'test this web app', 'QA', 'bug hunt', 'scrape page') with synonyms, giving comprehensive coverage of terms a user would actually say; file extensions are not applicable to this domain, so it still fits the level-5 anchor.	5 / 5
Distinctiveness Conflict Risk	The Open Design preview-validation framing carves a fairly distinct niche, but 'browser automation / testing web apps / QA / bug hunts' is a crowded space with minor overlap risk against other web/browser testing skills, fitting 'mostly distinct; minor overlap risk' (level 4) rather than the minimal-conflict level 5.	4 / 5
	Total	19 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: nexu-io/open-design
Path: skills/agent-browser/SKILL.md
Commit: 1a7a23d

Reviewed: about 7 hours 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.