Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/Encod3d-Sec/TORCHnpx agentmods add skills/encod3d-sec/torch/chrome-devtools-browserWrote 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/encod3d-sec/torch/chrome-devtools-browser)<a href="https://agentmods.dev/skills/encod3d-sec/torch/chrome-devtools-browser"><img src="https://agentmods.dev/badge/skills/encod3d-sec/torch/chrome-devtools-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/encod3d-sec/torch/chrome-devtools-browser"><img src="https://agentmods.dev/badge/skills/encod3d-sec/torch/chrome-devtools-browser.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00162 | $0.01254 |
| Opus 5.5 | $0.00065 | $0.00502 |
| Sonnet 5.5 | $0.00032 | $0.00251 |
| Haiku 4.5 | $0.00016 | $0.00125 |
Grade A, and why
chrome-devtools-browser scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
| **The real API calls the app makes** | `list_network_requests` (add `includePreservedRequests:true` to span the login redirects); this is the authenticated API map a curl crawl never sees | How it starts
The opening of the file, as written. The whole thing — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
chrome-devtools-browser
Headless browser.sh cannot solve a human login. This skill brings up a visible chromium on the
Kali desktop (:0) the operator can click and type into, with its DevTools port tunnelled back so the
chrome-devtools MCP drives and watches it. Two roles at once: the human authenticates; the agent
observes + acts. The VM holds the VPN/egress path, so it reaches both internet and in-scope targets.
1. Bring it up (one command)
bash scripts/browser-visible.sh <login-url> --profile bbtest-<eng>
# e.g. bash scripts/browser-visible.sh https://target.example/ib --profile bbtest-<eng>
It resolves the VM's seat session, unlocks/wakes :0, frees the CDP port, launches chromium as the
seat user on :0 in a named tmux session (a dedicated per-engagement profile, so the operator's own
profile/cookies stay clean), then reuses scripts/browser.sh to forward CDP to http://127.0.0.1:9222.
It prints VISIBLE on :0 when a real on-screen window exists (not merely a listening port). If it prints
a no (:N) desktop session message, the VM has no desktop - fall back to scripts/browser.sh (headless).
2. Attach + hand off for login
- The chrome-devtools MCP attaches to
http://127.0.0.1:9222. Confirm withlist_pages;navigate_pageto the login URL if needed. - Stay off the browser while the operator enters credentials. Take a
take_screenshotso they can confirm the page, then wait for their "logged in" before you drive again. The operator owns credential entry and the phone approval; the agent never types a personal code / PIN.
3. Drive + observe (chrome-devtools MCP capability map)
| Need | MCP tool |
|---|---|
| List/select tabs | list_pages (select the target tab; ignore the operator's other tabs) |
| Go to a page / back / reload | navigate_page |
| Rendered DOM (elements + uids) | take_snapshot |
| Visual PoC / confirm state | take_screenshot -> a web evidence image (hand to Skill(screenshot)/Skill(evidence)) |
| The real API calls the app makes | list_network_requests (add includePreservedRequests:true to span the login redirects); this is the authenticated API map a curl crawl never sees |
| One request's headers/body/cookies | get_network_request <reqid> |
| Console / JS errors | list_console_messages |
| Read client config / globals, confirm DOM-XSS | evaluate_script (e.g. __NEXT_DATA__, window.*, fire a payload in the real DOM) |
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 · 76 lines · 162 tokens per session scan A 062f979560ce
chrome-devtools-browser is a skill published in the GitHub repository Encod3d-Sec/TORCH (336 stars, last pushed 1mo ago), licensed MIT. It adds 162 tokens to every session and 1,254 once invoked, about $0.0006 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
debug-optimize-lcp
Guides debugging and optimizing Largest Contentful Paint (LCP) using Chrome DevTools MCP tools. Use this skill whenever the user asks about LCP performance, slow page loads, Core Web Vitals optimization, or wants to understand why their page's main content takes too long to appear. Also use when the user mentions…
interactive-login
Log a CLI in with a browser/device-code flow (aws sso, gh, glab, gcloud, …) and save the result to the keychain so later commands can request it through execute.credentials.
pinchtab-mcp
Use this skill when a task requires browser automation through PinchTab's MCP server connected to a remote browser instance. Covers navigation, element interaction, data extraction, form filling, multi-step flows, and session management via MCP tools.
azure-messaging-webpubsub-java
Build real-time web applications with Azure Web PubSub SDK for Java. Use when implementing WebSocket-based messaging, live updates, chat applications, or server-to-client push notifications.
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
open-design-browser
Inspect and exercise local HTML design artifacts and library prototype previews with the existing agent-browser CLI, including screenshots, interaction checks, and rendered DOM snapshots. Use when rendered evidence helps author or review a design; excludes browser installation, remote services, and unrelated browser…