CtrlK
BlogDocsLog inGet started
Tessl Logo

agent-browser

Drive a real browser to inspect or interact with a web page or app — navigate, take screenshots, read console and network, fill simple forms — for verification tasks, not unattended automation.

57

Quality

66%

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

Fix and improve this skill with Tessl

tessl review fix ./packages/skills-catalog/catalog/optional/browser/agent-browser/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

71%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 concise, well-structured guide that avoids over-explanation and clearly scopes supervised use, anti-patterns, and failure modes. Its main gap is the absence of any executable launch/drive code, which limits actionability for a code-oriented skill.

Suggestions

Add a short, copy-paste-ready launch snippet (e.g., Playwright or Puppeteer) showing the real-looking user agent, viewport, and wait-for-selector pattern so the skill is executable, not just descriptive.

Include an explicit screenshot/console-capture code example under 'What evidence to record' to make evidence collection directly actionable.

Tighten the validation loop in 'Driving the browser' by adding a verify-after-wait checkpoint before interaction, to push workflow clarity toward the explicit validate→fix→retry anchor.

DimensionReasoningScore

Conciseness

Lean throughout; assumes Claude knows what browsers and selectors are, with no padded concept explanations; every section earns its place and reads as terse, actionable guidance.

5 / 5

Actionability

Gives concrete procedural guidance and specific signals (e.g., viewport sizes, network-idle waits, 30s timeout) but contains no executable code or commands to launch/drive a browser; it instructs via prose rather than copy-paste-ready snippets.

3 / 5

Workflow Clarity

The 'Driving the browser' section is a clearly numbered sequence with wait/capture checkpoints and a 'close cleanly' end step, plus a retry-once escalation pattern; minor validation gaps keep it just below the explicit validate→fix→retry 5-anchor.

4 / 5

Progressive Disclosure

No bundle files are present and none are referenced; the body is a single, well-organized overview under ~80 lines with clear section headers, so for a self-contained skill this satisfies the simple-skill exception for a top score.

5 / 5

Total

17

/

20

Passed

Description

62%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 concrete and well-scoped to supervised browser verification, with a useful 'not unattended automation' boundary. It would benefit from explicit 'Use when...' trigger phrasing and more natural user-facing trigger terms.

Suggestions

Add an explicit 'Use when...' clause naming concrete triggers (e.g., 'Use when you need to screenshot a deployed page, reproduce a UI bug interactively, or read JS-rendered content that curl cannot see').

Surface natural trigger terms users actually say — 'screenshot a web page', 'console errors', 'puppeteer/playwright' — rather than only the generic 'web page or app'.

Consider noting file extensions or signals (e.g., 'SPAs', 'login-gated pages') to sharpen trigger matching.

DimensionReasoningScore

Specificity

Names the domain and several concrete actions ('navigate, take screenshots, read console and network, fill simple forms') with only minor gaps; falls just short of the comprehensive 5-anchor coverage.

4 / 5

Completeness

Clearly answers 'what' (drive a browser to inspect/interact) and 'when' ('for verification tasks, not unattended automation'), with the 'when' somewhat soft rather than an explicit 'Use when...' trigger clause.

4 / 5

Trigger Term Quality

Includes some relevant terms ('browser', 'web page or app', 'verification tasks') but misses the natural phrases a user would say like 'screenshot a page', 'puppeteer', or 'console errors'; mostly one generic cluster.

3 / 5

Distinctiveness Conflict Risk

Scoped to supervised browser verification and explicitly excludes scraping, giving a clear niche with only minor overlap risk against general HTTP-fetch skills.

4 / 5

Total

15

/

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.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository
paperclipai/paperclip
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.