CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

pixelbrowse

Screenshot and visually read any web page or document using pixelshot. Use instead of fetching raw HTML when you need to see what a page looks like, read visual content (charts, diagrams, infographics), check layouts, or verify UI. Triggers: "look at this page", "screenshot", "what does this site look like", "check the UI", "read this visually", "view this URL", viewing web content.

Quality

91%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

88%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 high-quality, actionable skill body: executable commands throughout, a validated three-step workflow with explicit failure modes and recovery guidance, and genuinely novel content (tile sizing for vision models, untrusted-tile handling, CDP reuse). The only improvements are minor tightening and possibly moving the advanced CDP section to a reference file.

DimensionReasoningScore

Conciseness

The body is efficient and tool-specific (flag values, tile-height rationale, manifest fields) with almost nothing Claude already knows, matching the 4 anchor: minor instances could be trimmed, e.g. the CDP prose loop and 'Pillow is always available (it's a pixelshot dependency)'. It is not the 5 anchor because a few sentences explain rather than instruct.

4 / 5

Actionability

Every command is copy-paste executable with concrete flags: pixelshot invocations for URLs, parallel URLs, wider viewports, and PDFs, plus a complete argv-based Python crop snippet; output path patterns and a real manifest example cover the common cases, matching the 5 anchor.

5 / 5

Workflow Clarity

The 3-step workflow (capture → read tiles.json → read every tile) includes an explicit validation checkpoint (the manifest), failure-mode handling ('complete: false — the capture is not trustworthy', 'No tiles.json at all — that is an error to report'), and feedback loops (re-running with --wait-network-idle or a different --viewport-width often fixes it), matching the 5 anchor; the destructive/batch cap does not apply since captures are read-only.

5 / 5

Progressive Disclosure

The single file has clear, well-organized sections with no nested or buried references and everything at one level, but at ~125 lines with advanced material (CDP attach setup, injection-hardened crop snippet) inlined rather than split into a reference file, it sits between the 4 anchor (good structure, minor organization gaps) and the 5 anchor's appropriately-split overview.

4 / 5

Total

18

/

20

Passed

Description

91%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 that clearly states capability, purpose, and explicit trigger phrases with a useful differentiator against raw HTML fetching. Its only deductions are the second-person 'when you need to' phrasing and mild overlap risk from broad terms like 'viewing web content'.

DimensionReasoningScore

Specificity

Lists multiple concrete actions ('Screenshot and visually read any web page or document', 'read visual content (charts, diagrams, infographics)', 'check layouts, or verify UI') which matches the 5 anchor, but the second-person phrasing 'when you need to see what a page looks like' triggers the rubric's third-person voice penalty, reducing the score by 1.

4 / 5

Completeness

Explicitly answers both: what ('Screenshot and visually read any web page or document using pixelshot') and when ('Use instead of fetching raw HTML when you need to see what a page looks like... Triggers: ...') with concrete trigger phrases, exactly matching the 5 anchor's example pattern.

5 / 5

Trigger Term Quality

Comprehensive natural trigger coverage with synonyms users would actually say: "look at this page", "screenshot", "what does this site look like", "check the UI", "read this visually", "view this URL" — matching the 5 anchor's breadth including variations; not the 4 anchor since no common natural phrasing is conspicuously missing.

5 / 5

Distinctiveness Conflict Risk

The skill carves a clear niche ('Use instead of fetching raw HTML', pixel-based visual reading) but 'screenshot' and 'viewing web content' could overlap other browsing/screenshot skills — fitting the 4 anchor (mostly distinct, minor overlap risk) rather than the 5 anchor's minimal conflict risk.

4 / 5

Total

18

/

20

Passed

Validation

93%

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

CriteriaDescriptionResult

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

Total

15

/

16

Passed

Repository

StarTrail-org/PixelRAG

Path

plugin/skills/pixelbrowse/SKILL.md

Reviewed

Table of Contents

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.