agent-browser

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

4.88x

Quality

87%

Does it follow best practices?

Impact

88%

4.88x

Average score across 3 eval scenarios

Securityby

High

Do not use without reviewing

Quality

Content

75%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 content is highly actionable with executable commands and a clear core workflow, but it underuses progressive disclosure: a substantial command reference and several pattern sections are inlined in SKILL.md while the referenced references/ and templates/ bundles are missing from the repo. Tightening the inline command catalog and ensuring referenced files exist would lift conciseness and progressive disclosure.

Suggestions

Move the full Essential Commands catalog and the longer Common Patterns into the (currently absent) references/ files, keeping only the core workflow and a few key examples inline in SKILL.md.

Create the referenced references/*.md and templates/*.sh files so the Deep-Dive Documentation and Ready-to-Use Templates tables point to real, navigable materials.

Add an explicit validation/checkpoint step (e.g. re-snapshot and assert expected element or URL) after form submissions and navigation to strengthen the feedback loop for batch/automated workflows.

Dimension	Reasoning	Score
Conciseness	Mostly lean and efficient with copy-pasteable commands and minimal concept explanation, but several overlapping command-list sections (Essential Commands plus multiple Common Patterns) repeat the same primitives and add some bulk that could be consolidated into references.	4 / 5
Actionability	Fully executable, copy-paste-ready bash commands throughout, with concrete annotated examples (snapshot output, refs) covering the common navigation/interaction/extraction cases.	5 / 5
Workflow Clarity	Clear four-step core workflow (Navigate -> Snapshot -> Interact -> Re-snapshot) with an explicit ref-lifecycle re-snapshot rule acting as a checkpoint, though there is no formal validate-after-action feedback loop for batch/destructive-style operations.	4 / 5
Progressive Disclosure	The body inlines a large command reference and many pattern sections rather than placing them in separate files, and the linked references/templates directories do not exist on disk, so the referenced deeper-dive materials are not actually present.	2 / 5
	Total	15 / 20 Passed

Description

100%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 is a strong model of the form: it states concrete capabilities, uses an explicit 'Use when' clause, and enumerates natural trigger phrases users would actually say. It is concise yet comprehensive with no padding or over-claims. The only minor note is that voice leans slightly toward feature-list rather than strict third-person action verbs, but it stays clean of first/second person.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions (navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps) with comprehensive coverage of the browser automation domain.	5 / 5
Completeness	Explicitly answers both 'what' (browser automation CLI for AI agents, with concrete capabilities) and 'when' (an explicit 'Use when...' clause plus enumerated 'Triggers include...' phrases).	5 / 5
Trigger Term Quality	Comprehensive coverage of natural trigger phrases users would say ('open a website', 'fill out a form', 'click a button', 'take a screenshot', 'scrape data from a page', 'test this web app', 'login to a site') with multiple synonyms included.	5 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (programmatic web interaction via browser automation) with distinct, specific triggers that are unlikely to fire for unrelated skills.	5 / 5
	Total	20 / 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
relative_links	Relative link issues: 9 missing	Warning
referenced_paths_exist	Referenced path issues: 12 missing	Warning

	Total	14 / 16 Passed

Repository: shanraisshan/claude-code-best-practice
Path: .claude/skills/agent-browser/SKILL.md
Commit: d4df0ac

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.