Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add denenis-lab/my-claude-skills --skill browsergit clone --depth 1 https://github.com/denenis-lab/my-claude-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/denenis-lab/my-claude-skills/browser)<a href="https://agentmods.dev/skills/denenis-lab/my-claude-skills/browser"><img src="https://agentmods.dev/badge/skills/denenis-lab/my-claude-skills/browser/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/denenis-lab/my-claude-skills/browser"><img src="https://agentmods.dev/badge/skills/denenis-lab/my-claude-skills/browser.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00053 | $0.01111 |
| Opus 5 | $0.00026 | $0.00556 |
| Sonnet 5 | $0.00011 | $0.00222 |
| Haiku 4.5 | $0.00005 | $0.00111 |
Grade C, and why
browser scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
curl -s http://localhost:9222/json/list | python3 -c " Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -s --max-time 2 http://localhost:9222/json/version How it starts
The opening of the file, as written. The whole thing — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Browser Skill
Connects to Chrome Canary via CDP (Chrome DevTools Protocol) using an isolated bot profile.
IMPORTANT: The user's main Chrome is never touched. Canary runs in parallel with a separate profile.
When to invoke
- User says "open in browser", "browse to", "check in Chrome", "what tabs are open", etc.
- Need to read web page content
- Need to open a URL
Connection algorithm
1. Check if CDP is already running
curl -s --max-time 2 http://localhost:9222/json/version
- If response → CDP is active, go to step 3
- If no response → go to step 2
2. Launch Chrome Canary with bot profile
2a. Check if port is free
lsof -i :9222 2>/dev/null && echo "PORT_BUSY" || echo "PORT_FREE"
If port is busy — investigate what's using it.
2b. Launch Canary
Bot profile is stored in ~/.chromium-bot. It persists cookies and logins between sessions.
"/Applications/Google Chrome Canary.app/Contents/MacOS/Google Chrome Canary" \
--remote-debugging-port=9222 \
--user-data-dir="$HOME/.chromium-bot" \
--no-first-run &
sleep 5
Verify connection:
curl -s --max-time 2 http://localhost:9222/json/version
If no response — wait 3 more seconds and retry.
3. List open tabs
curl -s http://localhost:9222/json/list | python3 -c "
import json, sys
tabs = json.load(sys.stdin)
pages = [t for t in tabs if t['type'] == 'page']
if not pages:
print('No open tabs')
else:
for i, t in enumerate(pages, 1):
print(f\"{i}. {t['title']}\n {t['url']}\n\")
"
4. Open a URL
curl -s -X PUT "http://localhost:9222/json/new?URL_HERE"
5. Read page content
5a. Ensure websockets is installed
python3 -c "import websockets" 2>/dev/null || pip3 install --break-system-packages websockets
5b. Read content
python3 -c "
import json, asyncio, websockets
async def get_content(ws_url):
async with websockets.connect(ws_url) as ws:
await ws.send(json.dumps({
'id': 1,
'method': 'Runtime.evaluate',
'params': {'expression': 'document.body.innerText'}
}))
result = json.loads(await ws.recv())
text = result.get('result', {}).get('result', {}).get('value', 'Could not read page')
print(text[:15000])
import urllib.request
tabs = json.loads(urllib.request.urlopen('http://localhost:9222/json/list').read())
pages = [t for t in tabs if t['type'] == 'page']
if pages:
asyncio.run(get_content(pages[0]['webSocketDebuggerUrl']))
else:
print('No open tabs')
"
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 153 lines · 53 tokens per session scan C 96cb631143b5
browser is a skill published in the GitHub repository denenis-lab/my-claude-skills (2 stars, last pushed 2mo ago), licensed MIT. It adds 53 tokens to every session and 1,111 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
js-reverse-automation
A workflow for examining JavaScript in a real browser to find where login data, API requests, form fields, or responses are encrypted or signed. It then produces code and connection documentation for using that logic elsewhere.
yao-chatgpt-crawler
Use when a user provides ChatGPT web AI-search keywords, repeat count, target entity, entity type, OpenCLI profile, and crawl interval preference, then needs repeated crawls aggregated into JSON plus a Kami HTML GEO report. Not for generic crawling, ChatGPT API chat, SEO writing, or one-off answers.
yao-doubao-crawler
Use when a user needs repeated Doubao AI-search collection from web or Android Appium into compatible JSON plus Markdown/Excel/HTML GEO reports. Requires keywords/questions and repeat count; target entity/type are required only for target-vs-competitor diagnosis. Not for generic scraping, Doubao API chat, hidden API…
browser-harness-authoring
Use when mapping a repeatable website workflow into a verified Hermes skill so later runs can follow known steps instead of rediscovering the site. Surveys browser compatibility, semantic targets, failure modes, recovery paths, decision gates, and expiry using dummy data and no irreversible submissions.
reddit-browse-and-post
Let an agent read Reddit by default, and publish only after the user explicitly opts in and approves the exact post. Everything here is account-agnostic: the user supplies credentials through environment variables or a browser login, never through chat.
openclaw-browser-auto
A setup guide for connecting OpenClaw, an AI-agent platform, to a remote Chrome browser through CDP, a browser control interface. It covers Docker-hosted browsers and Browserless.io, a hosted browser service.