dev-browser

Fallback browser automation with persistent Chrome state. Use only when Browser Use is unavailable or blocked.

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Fix and improve this skill with Tessl

tessl review fix ./.agents/skills/dev-browser/SKILL.md

Quality

Content

73%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 with executable commands, explicit validation checkpoints, and error-recovery loops, and it is well-organized into clear sections. Its main weakness is significant redundancy — the same profile/isolation rules are restated across Plate Defaults, Rules, and Google Chrome Path — which inflates token cost without adding guidance.

Suggestions

Merge the Rules section into Plate Defaults (or vice versa) — the dedicated --user-data-dir, separate-profile, and macOS open -na rules are each stated two to three times across sections.

Move the ~40-line Google Chrome Path profile-cloning procedure (Local State mapping, rsync clone, launch) into a references/ file, keeping SKILL.md as a lean overview with a one-level-deep pointer.

State the non-macOS (Linux) launch alternative or explicitly scope the skill to macOS so the Google Chrome Path sequence does not read as incomplete on other platforms.

Dimension	Reasoning	Score
Conciseness	Several rules are repeated nearly verbatim across sections — the dedicated --user-data-dir / don't-use-the-daily-Default-profile rule appears in Plate Defaults, Rules (lines 45-48), Preferred Shape, and Google Chrome Path, and the macOS 'open -na' guidance appears twice (lines 31 and 49) — making the Rules section largely a padded restatement of Plate Defaults. Not score 3 because this is whole-section redundancy rather than minor tightening; not score 1 because there is no explanation of concepts Claude already knows and the remaining content is dense and useful.	2 / 5
Actionability	Everything is copy-paste executable: install commands, curl sanity check with expected healthy output, exact rsync clone with exclusions, python profile-mapping snippet, heredoc connect/verify scripts, and lsof to identify a stale listener, covering the common cases end to end.	5 / 5
Workflow Clarity	The Google Chrome Path section gives a numbered five-step sequence with concrete commands, and validation is explicit throughout: /json/version health check with expected output, a page.title() verification, a websocket-URL fallback when CDP resolution fails, and an error-recovery loop (lsof the listener, kill it, relaunch) — matching the anchor for clear sequence with explicit validation steps and feedback loops.	5 / 5
Progressive Disclosure	No bundle files exist and the single SKILL.md is well-sectioned with clear headers (Installation, Plate Defaults, Fallback Setup, Rules, Preferred Shape, Google Chrome Path), so structure is good. Not 5 because the ~40-line Chrome profile-cloning setup could live in a separate reference file and the skill exceeds the under-50-lines simple-skill exception.	4 / 5
	Total	16 / 20 Passed

Description

66%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 has an explicit and well-scoped trigger clause that clearly distinguishes it from the primary browser skill, but the capability statement is thin — it names the domain and one characteristic without concrete actions or natural trigger variations. It is concise and in third person, avoiding fluff.

Suggestions

Add 1-2 concrete actions to the capability clause, e.g. 'Connects to a persistent Chrome debug profile over CDP on port 9222 and drives pages for automation tasks.'

Include natural trigger variations users would say, such as 'Chrome debugging', 'CDP', or 'browser automation fallback', alongside 'Browser Use is unavailable or blocked'.

Briefly state the outcome or scope of use (e.g. browsing, scraping, verifying web UIs) so the 'what' is answerable without opening the skill body.

Dimension	Reasoning	Score
Specificity	"Fallback browser automation with persistent Chrome state" names the domain and one concrete capability (persistent Chrome state) but lists no concrete actions such as connecting via CDP, reusing a debug profile, or driving pages, matching the anchor for domain plus 1-2 concrete actions without comprehensive coverage.	3 / 5
Completeness	Both parts are present: what ("Fallback browser automation with persistent Chrome state") and an explicit when ("Use only when Browser Use is unavailable or blocked"), but the what is thin and the when could be more specific, matching the both-present-but-could-be-more-explicit anchor rather than the fully concrete anchor 5.	4 / 5
Trigger Term Quality	Relevant keywords like "browser automation", "Chrome", and "Browser Use" are present, but common variations a user would naturally say (dev browser, Chrome debugging, CDP, Puppeteer/Playwright alternative) are missing, matching the "some relevant keywords but missing common variations" anchor.	3 / 5
Distinctiveness Conflict Risk	The "Use only when Browser Use is unavailable or blocked" clause carves out a clear fallback niche relative to a specific named plugin, giving distinct triggers and minimal conflict risk with other browser skills.	5 / 5
	Total	15 / 20 Passed

Validation

81%

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 — 13 / 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
relative_links	Relative link issues: 1 suspicious	Warning

	Total	13 / 16 Passed

Repository: udecode/plate
Path: .agents/skills/dev-browser/SKILL.md
Commit: babb3c2

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.