ReddRadar
Agent skill
cmux-browser
End-user browser automation with cmux. Use when you need to open sites, interact with pages, wait for state changes, and extract data from cmux browser surfaces.
Install this agent skill to your Project
npx add-skill https://github.com/manaflow-ai/cmux/tree/main/skills/cmux-browser
SKILL.md
Browser Automation with cmux
Use this skill for browser tasks inside cmux webviews.
Core Workflow
- Open or target a browser surface.
- Verify navigation with
get urlbefore waiting or snapshotting. - Snapshot (
--interactive) to get fresh element refs. - Act with refs (
click,fill,type,select,press). - Wait for state changes.
- Re-snapshot after DOM/navigation changes.
cmux --json browser open https://example.com
# use returned surface ref, for example: surface:7
cmux browser surface:7 get url
cmux browser surface:7 wait --load-state complete --timeout-ms 15000
cmux browser surface:7 snapshot --interactive
cmux browser surface:7 fill e1 "hello"
cmux --json browser surface:7 click e2 --snapshot-after
cmux browser surface:7 snapshot --interactive
Surface Targeting
# identify current context
cmux identify --json
# open routed to a specific topology target
cmux browser open https://example.com --workspace workspace:2 --window window:1 --json
Notes:
- CLI output defaults to short refs (
surface:N,pane:N,workspace:N,window:N). - UUIDs are still accepted on input; only request UUID output when needed (
--id-format uuids|both). - Keep using one
surface:Nper task unless you intentionally switch.
Wait Support
cmux supports wait patterns similar to agent-browser:
cmux browser <surface> wait --selector "#ready" --timeout-ms 10000
cmux browser <surface> wait --text "Success" --timeout-ms 10000
cmux browser <surface> wait --url-contains "/dashboard" --timeout-ms 10000
cmux browser <surface> wait --load-state complete --timeout-ms 15000
cmux browser <surface> wait --function "document.readyState === 'complete'" --timeout-ms 10000
Common Flows
Form Submit
cmux --json browser open https://example.com/signup
cmux browser surface:7 get url
cmux browser surface:7 wait --load-state complete --timeout-ms 15000
cmux browser surface:7 snapshot --interactive
cmux browser surface:7 fill e1 "Jane Doe"
cmux browser surface:7 fill e2 "jane@example.com"
cmux --json browser surface:7 click e3 --snapshot-after
cmux browser surface:7 wait --url-contains "/welcome" --timeout-ms 15000
cmux browser surface:7 snapshot --interactive
Clear an Input
cmux browser surface:7 fill e11 "" --snapshot-after --json
cmux browser surface:7 get value e11 --json
Stable Agent Loop (Recommended)
# navigate -> verify -> wait -> snapshot -> action -> snapshot
cmux browser surface:7 get url
cmux browser surface:7 wait --load-state complete --timeout-ms 15000
cmux browser surface:7 snapshot --interactive
cmux --json browser surface:7 click e5 --snapshot-after
cmux browser surface:7 snapshot --interactive
If get url is empty or about:blank, navigate first instead of waiting on load state.
Deep-Dive References
| Reference | When to Use |
|---|---|
| references/commands.md | Full browser command mapping and quick syntax |
| references/snapshot-refs.md | Ref lifecycle and stale-ref troubleshooting |
| references/authentication.md | Login/OAuth/2FA patterns and state save/load |
| references/authentication.md#saving-authentication-state | Save authenticated state right after login |
| references/session-management.md | Multi-surface isolation and state persistence patterns |
| references/video-recording.md | Current recording status and practical alternatives |
| references/proxy-support.md | Proxy behavior in WKWebView and workarounds |
Ready-to-Use Templates
| Template | Description |
|---|---|
| templates/form-automation.sh | Snapshot/ref form fill loop |
| templates/authenticated-session.sh | Login once, save/load state |
| templates/capture-workflow.sh | Navigate + capture snapshots/screenshots |
Limits (WKWebView)
These commands currently return not_supported because they rely on Chrome/CDP-only APIs not exposed by WKWebView:
- viewport emulation
- offline emulation
- trace/screencast recording
- network route interception/mocking
- low-level raw input injection
Use supported high-level commands (click, fill, press, scroll, wait, snapshot) instead.
Troubleshooting
js_error on snapshot --interactive or eval
Some complex pages can reject or break the JavaScript used for rich snapshots and ad-hoc evaluation.
Recovery steps:
cmux browser surface:7 get url
cmux browser surface:7 get text body
cmux browser surface:7 get html body
- Use
get urlfirst so you know whether the page actually navigated. - Fall back to
get text bodyorget html bodywhensnapshot --interactiveorevalreturnsjs_error. - If the page is still failing, navigate to a simpler intermediate page, then retry the task from there.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
manaflow-ai/cmux
release
Prepare and ship a cmux release end-to-end: choose the next version, curate user-facing changelog entries, bump versions, open and monitor a release PR, merge, tag, and verify published artifacts. Use when asked to cut, prepare, publish, or tag a new release.
manaflow-ai/cmux
cmux-debug-windows
Manage cmux debug windows and related debug menu wiring for Sidebar Debug, Background Debug, and Menu Bar Extra Debug. Use this when the user asks to open/tune these debug controls, add or adjust Debug menu entries, or capture/copy a combined debug config snapshot.
manaflow-ai/cmux
cmux-markdown
Open markdown files in a formatted viewer panel with live reload. Use when you need to display plans, documentation, or notes alongside the terminal with rich rendering (headings, code blocks, tables, lists).
manaflow-ai/cmux
cmux
End-user control of cmux topology and routing (windows, workspaces, panes/surfaces, focus, moves, reorder, identify, trigger flash). Use when automation needs deterministic placement and navigation in a multi-pane cmux layout.
davila7/claude-code-templates
verl-rl-training
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
davila7/claude-code-templates
openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
Didn't find tool you were looking for?