browser
Web browser automation with AI-optimized snapshots for claude-flow agents
59
51%
Does it follow best practices?
Impact
57%
2.52xAverage score across 3 eval scenarios
Risky
Do not use without reviewing
Optimize this skill with Tessl
npx tessl skill review --optimize ./.claude/skills/browser/SKILL.mdEvaluation results
41%
15%Tech News Headline Tracker
Snapshot optimization and element refs
Uses agent-browser CLI
100%
100%
Opens URL with agent-browser
100%
100%
Snapshot with -i flag
0%
0%
Element refs used for interaction
0%
0%
Re-snapshot after navigation
0%
0%
Core 4-step workflow order
0%
0%
Data saved to file
100%
100%
Uses get text for extraction
0%
0%
No CSS selector fallback
0%
100%
Browser closed at end
0%
100%
74%
67%Open-Source Repository Comparison Tool
Session-based parallel browsing
Uses --session flag
0%
100%
Multiple named sessions
0%
100%
Snapshot -i per session
0%
20%
Element refs in sessions
0%
0%
Parallel or sequential session operations
0%
100%
Session open steps
0%
100%
Re-snapshot within session
0%
0%
Results saved to file
100%
100%
No plain agent-browser without session
0%
100%
Browser sessions closed
0%
100%
60%
23%Automated GitHub Account Search Flow
Sequential interaction with wait handling
Uses agent-browser fill
87%
0%
Uses snapshot -i before each interaction
0%
41%
Element refs for all form interactions
0%
50%
Wait after navigation
86%
100%
Re-snapshot after step transition
0%
16%
Click uses refs
12%
100%
Core workflow order
30%
40%
Wait type is appropriate
87%
100%
Verification step
75%
62%
Browser closed at end
0%
100%
- Repository
- ruvnet/ruvector
- Commit
462536e
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
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.