Name: browser
Rating: 67.60000000000001 (1 reviews)
Author: ruvnet

Blog Docs Log inGet started

browser

Web browser automation with AI-optimized snapshots for claude-flow agents

2.52x

Quality

63%

Does it follow best practices?

Impact

57%

2.52x

Average score across 3 eval scenarios

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/browser/SKILL.md

The canonical home for this skill is browser in ruvnet/ruv-FANN

Quality

Content

78%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: executable commands throughout, four complete worked examples, and a clear core workflow with implicit feedback via re-snapshotting and wait conditions. Its main weaknesses are the absence of any progressive disclosure (the entire command reference is inlined in SKILL.md with no reference files) and missing explicit error-recovery checkpoints in the workflow.

Suggestions

Move the Quick Reference command tables to a references/ file (e.g. references/commands.md) and keep only the core workflow and a few key commands in SKILL.md.

Add an explicit error-recovery step to the Core Workflow, e.g. what to do when a click fails or an element ref becomes stale (re-snapshot and retry).

Trim the 'Integration with Claude Flow' section (MCP tools list, memory, hooks) to the essentials or relocate it to a reference file.

Dimension	Reasoning	Score
Conciseness	The body is dense and command-focused — mostly tables and worked examples with no explanations of concepts Claude already knows. It is not 5 because sections like 'Integration with Claude Flow' (MCP tools, memory, hooks) add tangential material that could be trimmed, and the full command reference pushes token budget.	4 / 5
Actionability	Every section provides copy-paste-ready executable commands, and the examples (login flow, form submission, data extraction, multi-session) cover the common use cases end-to-end including wait conditions. Fully matches the top anchor.	5 / 5
Workflow Clarity	The Core Workflow gives a clear numbered sequence (open → snapshot → interact → re-snapshot) with implicit checkpoints (re-snapshot after page changes, wait --url / wait --text in examples). It is not 5 because there is no explicit validation/error-recovery guidance for failed clicks, stale refs, or navigation errors.	4 / 5
Progressive Disclosure	The single SKILL.md is well-sectioned with one-level structure and no buried references, but ~60 lines of full command reference tables are inlined with no bundle files at all — matching 'Some structure but could be better organized; content that should be separate is inline'. It is not 2 because organization and signaling within the file are clear, not minimal.	3 / 5
	Total	16 / 20 Passed

Description

48%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 communicates a clear domain and a differentiating feature (AI-optimized snapshots) but reads more like a tagline than a capability description: it enumerates no concrete actions and provides no 'when to use' guidance, which caps completeness. Natural trigger terms users would actually say (scrape, navigate, screenshot) are absent from the description despite being present in the frontmatter triggers list.

Suggestions

Add an explicit 'Use when...' clause, e.g. 'Use when navigating websites, filling forms, taking screenshots, or scraping page content.'

Enumerate 3-4 concrete capabilities in the description (open pages, click/fill via element refs, capture screenshots, extract text) instead of only naming the domain.

Fold natural trigger words (browse, navigate, screenshot, scrape, form filling) into the description itself so users' phrasing matches it.

Dimension	Reasoning	Score
Specificity	The description names the domain ("Web browser automation") and one implementation detail ("AI-optimized snapshots") but lists no concrete actions such as navigate, click, fill, or extract — matching the anchor 'Names the domain but actions are minimal or generic'. It does not reach 3 because no 1-2 concrete user-facing actions are stated.	2 / 5
Completeness	The 'what' is clear (browser automation with AI-optimized snapshots), but there is no 'Use when...' clause or equivalent explicit trigger guidance, which caps completeness at 3 per the judging guidelines. It is not 2 because the 'what' is specific rather than vague.	3 / 5
Trigger Term Quality	"Web browser automation" is a relevant keyword, but the description omits natural phrases users would say such as scrape, navigate, screenshot, open a page, or form filling. It fits 'Some relevant keywords but missing common variations or synonyms' rather than 4, which would require broader coverage.	3 / 5
Distinctiveness Conflict Risk	Browser automation with agent-browser snapshots for claude-flow agents is a fairly distinct niche with minor overlap risk against general web-fetching/scraping skills. It is not 5 because 'web automation' is a broad phrase that could overlap with other web-interaction skills.	4 / 5
	Total	12 / 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: ruvnet/RuVector
Path: .claude/skills/browser/SKILL.md
Commit: 4b59352

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.