lens-inspect

Use the Lens built-in browser to inspect a running web application. Trigger when the request involves previewing a page, taking screenshots, reading DOM, checking console errors, picking elements to locate source code, or comparing rendered output to a design spec.

1.02x

Quality

81%

Does it follow best practices?

Impact

92%

1.02x

Average score across 2 eval scenarios

Securityby

Low

Low-risk findings worth noting

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 body is a well-organized, largely executable reference: a complete tool table, precise session semantics, and a numbered design-comparison loop with a built-in retry. The main gaps are a handful of user-doc tokens, no example selectors/outputs for the interaction tools, and detail that could move to a reference file as the skill grows.

Suggestions

Trim user-facing lines (Globe icon / Command Palette) that describe the UI rather than agent behavior, and fold the 'What is Lens?' section into a single sentence.

Add one concrete example per interaction tool, e.g., a sample CSS selector for `stave_lens_click` and expected box-model output from `stave_lens_inspect`.

If the skill grows, move the React `_debugSource` mechanics and picker data-field lists into a `references/` file and keep SKILL.md as the overview.

Dimension	Reasoning	Score
Conciseness	The body is lean — tool table, dense bullets, no padding explaining concepts Claude already knows — so it is clearly above the 'mostly efficient' midpoint. A few tokens are spent on user-facing documentation ('Users can also open Lens with the Globe icon or the Command Palette action Show Lens') and the two-sentence 'What is Lens?' section could be tightened, placing it at 'minor instances of over-explanation' rather than 'every token earns its place'.	4 / 5
Actionability	Concrete guidance dominates: all 14 MCP tools are listed with purposes, session rules reference exact tool names, and the Figma workflow includes a literal call ('stave_lens_navigate("http://localhost:3000/page")'). It falls short of fully copy-paste-ready 5 because click/type/inspect lack example selectors or expected output shapes.	4 / 5
Workflow Clarity	The Figma comparison workflow is clearly numbered with an explicit feedback loop ('Modify code to fix discrepancies... Repeat from step 2'), and the session presentation boundary rules are precise. Other guidance (check console/network logs early) is stated as tips rather than sequenced checkpoints, so it does not reach the 'explicit validation steps' bar of 5.	4 / 5
Progressive Disclosure	Sections are well-organized and appropriately scoped for a single-file skill, but at ~96 lines some detail (React _debugSource mechanics, picker captured-data fields) could be split into a reference file. With no bundle files present and content exceeding the 'under 50 lines' simple-skill exception, this lands at 'good structure; minor organization gaps' rather than a fully split 5.	4 / 5
	Total	16 / 20 Passed

Description

87%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: it names the tool and domain, gives an explicit trigger clause with six concrete natural-language capabilities, and stays clearly distinct from neighboring skills. The only improvement space is broader capability/keyword coverage (measuring, network logs, interaction terms).

Suggestions

Add one or two more capability keywords to the trigger list — e.g., measuring spacing between elements, reading network requests, or clicking/typing to interact with the page.

Include common user synonyms such as 'inspect the page', 'visual regression', or 'check the rendered UI' to broaden natural trigger coverage.

Dimension	Reasoning	Score
Specificity	The what-clause ('Use the Lens built-in browser to inspect a running web application') is thin on its own, but the trigger list names six concrete capabilities — 'previewing a page, taking screenshots, reading DOM, checking console errors, picking elements to locate source code, or comparing rendered output to a design spec'. Minor gaps in coverage (measuring, network logs, click/type) keep it below the comprehensive 5 anchor.	4 / 5
Completeness	It explicitly answers both what ('Use the Lens built-in browser to inspect a running web application') and when ('Trigger when the request involves...') with concrete trigger phrases, matching the 5 anchor exactly. The 4 anchor's 'when could be more explicit' does not apply since the trigger clause is fully enumerated.	5 / 5
Trigger Term Quality	'Taking screenshots', 'checking console errors', 'previewing a page', and 'comparing rendered output to a design spec' are natural user phrasings tied to the domain. A few common variants ('inspect the page', 'visual regression', 'debug the UI') are missing, so it fits 'good keyword coverage; a few natural terms missing' rather than the comprehensive 5 anchor.	4 / 5
Distinctiveness Conflict Risk	'Lens built-in browser' carves out a clear niche and the trigger phrases are scoped to it, giving minimal conflict risk. Screenshot/design-spec terms carry only minor overlap with generic browser or Figma skills, and the explicit tool qualifier distinguishes them.	5 / 5
	Total	18 / 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: sendbird/stave
Path: skills/lens-inspect/SKILL.md
Commit: ec11c51

Reviewed: 6 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.