lynx-recorder
Skill28 starsChanged 8 months agoApache-2.0View on GitHub
Record Lynx page interactions into a TestBench replay file using agent-lynx. Use when asked to start or stop a recording, capture page interactions for replay, or troubleshoot recording capture. For inspecting or editing an existing fixture, use lynx-recorder-fixture.
In this repository17 other files it gives its agents.
- AGENTS.mdAGENTS.md
- habitat-usageSkill
- lynx-a2uiSkill
- lynx-api-docsSkill
- lynx-check-css-supportSkill
- lynx-debug-info-remappingSkill
- lynx-devtoolSkill
- lynx-openuiSkill
- lynx-recorder-fixtureSkill
- lynx-tasm-codecSkill
- lynx-trace-analysisSkill
- lynx-trace-recordSkill
- lynx-typescriptSkill
- lynx-uiSkill
- reactlynx-best-practicesSkill
- rspeedy-bundle-sizeSkill
- vanilla-lynxSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 54
Start before opening the page: Run recorder start. Capture must begin *before* the target page loads so the recording includes loadTemplate.
- Very specific: line 57
Stop and save: Run recorder end --output ./my-recording.json.
- Very specific: line 63
The target must be a recorder build: If start reports Recording.start is not implemented, the app/engine was built without …
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Lynx Recorder
Record Lynx page interactions and produce a TestBench replay file. The saved
.json file may contain base64(zlib(JSON)); use lynx-recorder-fixture to inspect
or edit its contents.
agent-lynx CLI
This skill contains no scripts. Recording is the recorder subcommand of
the public agent-lynx CLI:
agent-lynx recorder start [options]
agent-lynx recorder end [options]The agent-lynx CLI is normally available on PATH. If
command -v agent-lynx fails, use pnpx, for example:
pnpx agent-lynx recorder start [options]The agent-lynx CLI on PATH and the pnpx command are the supported
execution forms.
This skill does not provide a local scripts/index.mjs entry.
Commands
recorder start
agent-lynx recorder start [--client <clientId>]-c, --client <clientId>(optional) — Target client. If omitted, the first available client is used.
recorder end
agent-lynx recorder end [--client <clientId>] [--output <path>]-c, --client <clientId>(optional) — Target client. If omitted, the first available client is used.-o, --output <path>(optional) — Output file or directory. If omitted, defaults to~/.lynx-devtool/files/lynxrecorder/recording-<clientId>-<timestamp>.json.
Workflow
- Start before opening the page. Run
recorder start. Capture must begin before the target page loads so the recording includesloadTemplate. - If
startreportsrestartRequired: true,enable_debug_modewas just enabled — restart the app and runrecorder startagain (see Notes). - Open and interact. Open (or reload) the target Lynx page and perform the interactions to capture.
- Stop and save. Run
recorder end --output ./my-recording.json. Report the absolute path(s) returned insavedFilesto the user.
Notes
- A replayable recording requires
loadTemplate. Open or reload the target page afterrecorder start. A recording started against an already-open page still captures interactions but cannot be replayed. - First-time
enable_debug_modemay need a restart.startenablesenable_debug_modeif it is off; when the platform persists that switch,startreturns{ success: false, restartRequired: true }. Restart the app, then runstartagain. - The target must be a recorder build. If
startreportsRecording.start is not implemented, the app/engine was built withoutENABLE_TESTBENCH_RECORDER(not a dev/recorder build). Use a recorder-capable build. - One client by default. Omitting
--clientpicks the first available client. When multiple devices or apps are connected, pass--clientexplicitly.
Output
end prints a single JSON object on success:
{
"success": true,
"message": "Recording ended successfully.",
"savedFiles": ["/abs/path/recording.json"],
"diagnostics": [
{
"file": "/abs/path/recording.json",
"fileSizeBytes": 12345,
"healthy": true,
"actions": 42,
"hasTemplate": true,
"functionDistribution": { "loadTemplate": 1, "SendTouchEvent": 12 },
"verdict": "Valid recording — includes template load and interaction data"
}
]
}diagnostics.verdict classification:
hasTemplate: true→ valid, replayable.- Touch events present but no
loadTemplate→ useful for analyzing interactions, not replayable. - Actions present but no
loadTemplate→ useful for inspecting JSB calls / data updates, not replayable. - 0 actions → empty recording; nothing was captured.
- Cannot parse → the file is not a valid TestBench recording.
Troubleshooting
| Symptom | Cause | Action |
start returns restartRequired: true | enable_debug_mode was just enabled and persisted | Restart the app, then run recorder start again |
start errors Recording.start is not implemented | Not a dev/recorder build (no ENABLE_TESTBENCH_RECORDER) | Use a recorder-capable build of the app/engine |
Recording has no loadTemplate (not replayable) | Page was already open, or not reloaded after start | Re-record: run start first, then open/reload the page |
| Empty recording (0 actions) | No page activity between start and end | Ensure the page is opened and interacted with during recording |
end times out | Device disconnected, or start was never called | Verify the connection; ensure start ran first |
| No available clients | Device not connected or app not running | Connect the device via USB and open the DevTool-enabled app |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

