Is the Browser Trace skill free?

Yes. The skill itself is free from mxyhi/ok-skills. SkillProof publishes the install command and an independent test verdict at no cost.

What is the Browser Trace SkillProof Score?

8.4/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 7/10, docs & honesty 4/5.

How do I install Browser Trace?

Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.

Can I use Browser Trace with Cursor, Copilot, Gemini CLI, Codex or other AI tools?

The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.

Browser Trace

Records a read-only CDP trace of a browser run and bisects it into per-page searchable buckets

by mxyhi · mxyhi/ok-skills

Tested · Works

Install — copy the command Download source ↗

Test report

Verdict: Tested · Works
Score: 8.4/10
Tested: Jul 21, 2026
Environment: Claude Code 2.x (agent harness)
Upstream re-checked: Aug 10, 2026 · 0fd691e

Cloned the skill, confirmed all 8 scripts + REFERENCE/EXAMPLES exist (spot-checked 3 scripts, HTTP 200). Security scan clean: the only child_process use spawns the local `browse` CLI via argument arrays (no shell string, no curl|sh, no base64, no secret exfil). Could not run the live-capture half (`browse` CLI not installed), but the bisector is pure Node stdlib, so I hand-built a 10-event synthetic CDP firehose across two pages and ran bisect-cdp.mjs + query.mjs for real: it correctly split into pages/000 (example.com) and pages/001 (app.example.com/login), attributed all 3 failures by page (network.failed + console.error to pid0, runtime.exception to pid1), and wrote per-page summary.json with domain rollups and network byType counts — matching the documented shape exactly. Verified the `browse` npm package (latest 0.9.5) is the Browserbase CLI the docs reference.

Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works

Installs cleanly 5/5
Triggers reliably 5/5
Output vs. baseline 7/10
Docs & honesty 4/5

What Browser Trace does

Attaches a second, read-only Chrome DevTools Protocol client to a browser automation session, records the full CDP firehose plus screenshots and DOM dumps to NDJSON, then slices the stream into per-page network/console/DOM/page JSONL buckets with a summary.json overview. Triggers when the user wants to debug a failed browser-automation run, audit network/console activity, attach a trace to an in-progress session, or feed per-page summaries back into an agent loop. Requires the `browse` (Browserbase) CLI for the live-capture half; the bisector and query tools run on the Node standard library alone.

How to install Browser Trace

git clone --depth 1 https://github.com/mxyhi/ok-skills.git /tmp/browser-trace-src
mkdir -p ~/.claude/skills
cp -R /tmp/browser-trace-src/browser-trace ~/.claude/skills/browser-trace
# Live capture requires the Browserbase CLI: npm install -g browse (provides `browse cdp`)
# Local runs need Chrome launched with --remote-debugging-port=9222; remote runs need BROWSERBASE_API_KEY
# Optional: jq for ad-hoc querying of the bisected JSONL files
# The bisect-cdp.mjs / query.mjs analysis scripts need only Node 18+ (no npm install, no browse CLI)

Skills live in ~/.claude/skills/ (global) or .claude/skills/ (per-project). Restart Claude Code after installing.

Commands — how to trigger Browser Trace

/browser-trace Records a read-only CDP trace of a browser run and bisects it into per-page searchable buckets

It also activates on plain-language prompts like these:

Capture a trace of this failing browser run
Audit network and console activity from the last run
Attach a trace to my in-progress automation session

Frequently asked questions

Is the Browser Trace skill free?: Yes. The skill itself is free from mxyhi/ok-skills. SkillProof publishes the install command and an independent test verdict at no cost.
Does Browser Trace work with Claude Code?: We tested it with Claude Code 2.x (agent harness) on Jul 21, 2026. Verdict: Tested · Works. Cloned the skill, confirmed all 8 scripts + REFERENCE/EXAMPLES exist (spot-checked 3 scripts, HTTP 200). Security scan clean: the only child_process use spawns the local `browse` CLI via argument arrays (no shell string, no curl|sh, no base64, no secret exfil). Could not run the live-capture half (`browse` CLI not installed), but the bisector is pure Node stdlib, so I hand-built a 10-event synthetic CDP firehose across two pages and ran bisect-cdp.mjs + query.mjs for real: it correctly split into pages/000 (example.com) and pages/001 (app.example.com/login), attributed all 3 failures by page (network.failed + console.error to pid0, runtime.exception to pid1), and wrote per-page summary.json with domain rollups and network byType counts — matching the documented shape exactly. Verified the `browse` npm package (latest 0.9.5) is the Browserbase CLI the docs reference.
What is the Browser Trace SkillProof Score?: 8.4/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 7/10, docs & honesty 4/5.
How do I install Browser Trace?: Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
Can I use Browser Trace with Cursor, Copilot, Gemini CLI, Codex or other AI tools?: The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.