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 IsaiaScope/ai --skill iso-browsergit clone --depth 1 https://github.com/IsaiaScope/aiWrote 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/isaiascope/ai/iso-browser)<a href="https://agentmods.dev/skills/isaiascope/ai/iso-browser"><img src="https://agentmods.dev/badge/skills/isaiascope/ai/iso-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/isaiascope/ai/iso-browser"><img src="https://agentmods.dev/badge/skills/isaiascope/ai/iso-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 | $0.00166 | $0.01269 |
| Opus 5.5 | $0.00066 | $0.00508 |
| Sonnet 5.5 | $0.00033 | $0.00254 |
| Haiku 4.5 | $0.00017 | $0.00127 |
Grade A, and why
iso-browser scanned grade A with 0 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 1mo 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.
iso-browser
Browser automation defaults to a throwaway profile, which is signed out of everything and therefore useless for the tasks people actually want automated. This skill connects to the browser the user is already using.
Invocation: /iso-browser [status | doctor | setup | launch]. Default status.
| Command | What it does |
|---|---|
status |
Prints the resolved config and the live state — Chrome running, CDP reachable, MCP entry. |
agents |
Lists the agents installed on this machine and which already know the browser server. |
doctor |
Checks every prerequisite and names the fix for each failure. Exit code is the failure count. |
setup |
Adds the connect flag to the MCP entry, backing the file up first. Requires an MCP-client restart. |
launch |
browser-url mode only: starts Chrome on the debug port with a dedicated profile. |
All four run scripts/browser.sh. This file describes the surface; it holds no logic.
The rule that matters
Never type credentials into an automated browser. Attach to a session the user already established. Signing in is theirs to do; driving the result is yours. Beyond being a boundary worth keeping, it does not work: Google refuses OAuth from a CDP-controlled browser with "This browser or app may not be secure." Attaching to an already-authenticated session never triggers it.
Connect modes
Set browser.connect in ~/.config/iso/iso.json.
| Mode | Uses the real profile? | Cost |
|---|---|---|
autoConnect (default) |
yes | Chrome 144+; user approves once at chrome://inspect/#remote-debugging |
browser-url |
no — Chrome 136+ forbids the default profile with a debug port | one login per site, in that profile |
isolated |
no | the built-in default; signed out of everything |
Prefer autoConnect. The other two exist for older Chrome and for throwaway work.
Any agent, not just one
agents reports every known agent config on the machine and whether each
already has the browser server. mcp_config defaults to auto, which prefers
the config that already knows the server and otherwise takes the first that
exists — so nothing needs configuring on a typical machine.
What ships with it
5 files 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.
- 1mo ago First seen · 104 lines · 166 tokens per session scan A 870dc4c0ae4a
iso-browser is a skill published in the GitHub repository IsaiaScope/ai (2 stars, last pushed 7d ago), licensed MIT. It adds 166 tokens to every session and 1,269 once invoked, about $0.0007 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-05.
Other skills, from other repositories
cross-browser-testing
A cross-browser test-coverage guide for selecting browsers, rendering engines, devices, and versions. A browser engine is the software that turns web code into a displayed page.
ui-test-cypress
Use this skill when you need to design Cypress e2e and component testing plans with commands, fixtures, network stubbing, and CI reporting; triggers include Cypress UI testing, UI automation testing, and ui-test-cypress.
ui-test-playwright
Use this skill when you need to design Playwright Test suites with fixtures, projects, traces, screenshots, API plus UI coverage, and CI reporting; triggers include Playwright UI testing, UI automation testing, and ui-test-playwright.
ui-test-puppeteer
Use this skill when you need to design Puppeteer automation for Chromium-driven checks, screenshots, PDFs, network interception, and CDP use cases; triggers include Puppeteer UI testing, UI automation testing, and ui-test-puppeteer.
ui-test-selenium
Use this skill when you need to design Selenium WebDriver UI automation plans with stable locators, waits, Page Objects, Grid, and CI execution; triggers include Selenium UI testing, UI automation testing, and ui-test-selenium.
ui-test-testcafe
Use this skill when you need to design TestCafe UI automation with fixtures, selectors, roles, browser matrix execution, and reports; triggers include TestCafe UI testing, UI automation testing, and ui-test-testcafe.