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 0xNyk/oneclaw --skill browser-logingit clone --depth 1 https://github.com/0xNyk/oneclawWrote 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/0xnyk/oneclaw/browser-login)<a href="https://agentmods.dev/skills/0xnyk/oneclaw/browser-login"><img src="https://agentmods.dev/badge/skills/0xnyk/oneclaw/browser-login/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/0xnyk/oneclaw/browser-login"><img src="https://agentmods.dev/badge/skills/0xnyk/oneclaw/browser-login.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 | $0.00000 | $0.00879 |
| Opus 5.5 | $0.00000 | $0.00352 |
| Sonnet 5.5 | $0.00000 | $0.00176 |
| Haiku 4.5 | $0.00000 | $0.00088 |
Grade C, and why
browser-login scanned grade C 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 2mo 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.
Harvests environment variableshighData exfiltration
Enumerating or grepping the environment for keys collects credentials unrelated to what the mod says it does.
- Always collect secrets using `secret_prompt`, then pass only the returned `secret://...` handle to `browser` actions. How it starts
The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Browser Login (Login Machine Loop)
Goal: log into an arbitrary website without provider-specific scripts by looping:
- observe current page
- classify what screen it is
- act using stable refs/semantic selectors
- repeat until logged in
Non-Negotiables (Security)
- Never ask the user to paste passwords/OTP into chat.
- Always collect secrets using
secret_prompt, then pass only the returnedsecret://...handle tobrowseractions. - Do not store credentials in memory or write them to disk.
- Prefer
browser snapshotrefs (@e1,@e2) and semanticbrowser findover brittle CSS selectors.
Setup
browsertool must be enabled and the target domain must be inbrowser.allowed_domains.- Start by opening the login URL (or the product URL that redirects to login).
Core Loop
Repeat up to ~20 iterations:
-
Observe:
- Call
browser snapshotwithinteractive_only=true,compact=true. - If the snapshot is insufficient (missing context), take
browser screenshot.
- Call
-
Classify the screen (pick one):
credential_form: there are input fields and a primary submit.email_first: email/username only with a next/continue.password_only: password only (email is remembered).choice_screen: account/workspace picker or auth method selector.blocked_modal: cookie banner / modal blocking clicks.mfa_otp: asks for a code (TOTP/SMS/email).magic_link: instructs user to check email for a link.captcha: challenge detected (cannot solve reliably inside this loop).logged_in: you can see app content / avatar / logout / dashboard.loading: spinners, disabled UI; wait and resnapshot.
-
Act (handlers):
blocked_modal
- Find the dismiss/accept button (often "Accept", "Agree", "Close", "X").
- Use
browser clickby ref orbrowser findby text/role, then resnapshot.
credential_form / email_first / password_only
- Identify required fields (email/username, password).
- For each required secret:
- Use
secret_prompt:kind="text"for email/usernamekind="password"for password
- Fill using
browser fillwherevalueis the returnedsecret://...handle.
- Use
- Submit:
- Click the primary submit button (ref or
browser findby role/text).
- Click the primary submit button (ref or
- If there is a second step (email -> password), resnapshot and repeat.
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.
- 2mo ago First seen · 95 lines · 0 tokens per session scan C 04a67aa24057
browser-login is a skill published in the GitHub repository 0xNyk/oneclaw (3 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 879 tokens. A static security scan graded it C with 1 finding (harvests environment variables). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
browser-login compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| browser-login0xNyk/oneclaw | Skill | 3 | 2mo ago | MIT | C | 0 | Claude Code, Codex |
| peekabookaivyy/kabot | Skill | 21 | 7mo ago | MIT | A | 17 | Claude Code, Codex |
| codebase-classificationgptme/gptme | Skill | 4.5k | yesterday | MIT | A | 16 | Claude Code, Codex |
| python-replgptme/gptme | Skill | 4.5k | yesterday | MIT | A | 15 | Claude Code, Codex |
| harbor-daytonaav/harbor | Skill | 3.3k | 6d ago | Apache-2.0 | A | 67 | Claude Code |
| web-accessopen-octo/octo-agent | Skill | 141 | 2d ago | MIT | A | 139 | Claude Code, Codex |
| browser-qadamianvtran/local-operator | Skill | 229 | yesterday | MIT | A | 45 | Claude Code, Codex |