CtrlK
BlogDocsLog inGet started
Tessl Logo

agent-browser

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

Install with Tessl CLI

npx tessl i github:brianlovin/claude-config --skill agent-browser
What are skills?

92

Does it follow best practices?

Validation for skill structure

SKILL.md
Review
Evals

Evaluation results

100%

27%

Automate a Contact Form Submission

Core browser workflow with ref-based interactions

Criteria
Without context
With context

snapshot -i used

0%

100%

Ref-based interactions

100%

100%

fill vs type

100%

100%

wait networkidle

0%

100%

Re-snapshot after action

50%

100%

Open command used

100%

100%

Close browser

100%

100%

Script documents commands

100%

100%

Correct open/snapshot order

100%

100%

No bare selectors in interactions

100%

100%

Without context: $0.4359 · 2m 16s · 30 turns · 32 in / 4,242 out tokens

With context: $0.9643 · 2m 42s · 46 turns · 55 in / 7,938 out tokens

96%

47%

Record a Demo of a Web Search Workflow

Demo recording with explore-first workflow

Criteria
Without context
With context

Explore before record

66%

100%

record start with .webm path

33%

100%

record stop issued

30%

100%

snapshot -i during recording

0%

100%

Ref-based interactions in recording

0%

100%

Wait used in recording

50%

100%

Exploration documented

87%

100%

Close browser at end

42%

42%

Script file produced

87%

100%

No premature recording

100%

100%

Without context: $0.9062 · 6m 17s · 34 turns · 41 in / 14,825 out tokens

With context: $1.2440 · 4m 44s · 53 turns · 303 in / 13,250 out tokens

88%

43%

Set Up Reusable Authentication State for a Test Suite

Authentication state persistence

Criteria
Without context
With context

state save used

13%

100%

state load used

26%

100%

wait --url pattern

0%

0%

snapshot -i for login form

0%

100%

fill for credentials

100%

100%

Ref-based login interaction

0%

100%

Post-load verification

90%

100%

close browser

100%

100%

Script documents both phases

100%

100%

State file named clearly

100%

100%

Without context: $0.8558 · 3m 55s · 37 turns · 42 in / 8,721 out tokens

With context: $1.3838 · 4m 28s · 62 turns · 70 in / 10,470 out tokens

Evaluated
Agent
Claude Code

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.