Agent skill

Cw Test

by adecarolis in adecarolis/wfweb

Run an end-to-end CW (morse) decode round-trip between two virtual wfweb rigs.

GPL-3.0Auto-check passedTesting & QA

Install Cw Test

skills CLI
$ npx skills add adecarolis/wfweb --skill cw-test -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install adecarolis/wfweb cw-test --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/adecarolis/wfweb.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/cw-test .claude/skills/cw-test && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
cw-test
GitHub stars
106
Token cost
~1.9k tokens
SKILL.md length
925 words
Files
1
Skills in repo
5
Repo updated
First seen
Licence
GPL-3.0

At a glance

Run an end-to-end CW (morse) decode round-trip between two virtual wfweb rigs.

  • Works in 3 steps: ./wfweb and… → Repo .mcp.json includes… → Load the playwright tools via…
  • The user asks to test CW
  • SKILL.md covers Prerequisites (verify before…, Boot the bench, Per-tab UI bring-up (do for… and QSO sequence (single…, plus 2 more sections
  • Calls make

What it does

Cw Test is an agent skill from adecarolis/wfweb. Run an end-to-end CW (morse) decode round-trip between two virtual wfweb rigs. Boots the testrig bench, drives both browsers via Playwright MCP, has each station key a known string with the keyer while the other decodes it with the built-in CW decoder, verifies the copy, screenshots both, and tears down. Use when the user asks to test CW, run a CW QSO, verify the CW decoder, or validate the keyer TX / decode RX path.

Its SKILL.md is about 1.9k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Testing & QA, covering Browser testing. It works with Model Context Protocol and Playwright. The repository describes itself as: Turn any Icom transceiver into a web-accessible station. Headless wfview fork with browser-based control, audio streaming, SSB, FreeDV, AX.25, CW, JS8 and FT8. The licence is GPL-3.0.

When your agent uses it

  • The user asks to test CW
  • Verify the CW decoder
  • Validate the keyer TX / decode RX path

Example prompts

  • “/cw-test”

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. ./wfweb and ./tools/virtualrig/virtualrig are built. If not: qmake wfweb.pro && make -j$(nproc) and cd tools/virtualrig && qmake && make…
  2. Repo .mcp.json includes --ignore-https-errors for the playwright MCP. Verify: pgrep -af @playwright/mcp must show --ignore-https-errors…
  3. Load the playwright tools via ToolSearch: select:mcpplaywrightbrowser_navigate,mcpplaywrightbrowser_snapshot,mcpplaywrightbrowser_click,mcp…

What it can do on your machine

Read from SKILL.md and the folder at commit 7cd5d2f. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • make

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Cw Test loads about 1.9k tokens when it runs. Until then it costs about 107 tokens; SKILL.md has 925 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~107
When it runs · the whole SKILL.md, loaded when a task matches
~1.9k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from adecarolis/wfweb at commit 7cd5d2f, republished under its GPL-3.0 licence (© adecarolis). 925 words, ~1,903 tokens.

Download SKILL.mdSave it as .claude/skills/cw-test/SKILL.md (or your agent's skills folder).
name
cw-test
description
Run an end-to-end CW (morse) decode round-trip between two virtual wfweb rigs. Boots the testrig bench, drives both browsers via Playwright MCP, has each station key a known string with the keyer while the other decodes it with the built-in CW decoder, verifies the copy, screenshots both, and tears down. Use when the user asks to test CW, run a CW QSO, verify the CW decoder, or validate the keyer TX / decode RX path.

Run an end-to-end CW decode round-trip between two virtual rigs: each station keys a known string (server-side keyer → virtualrig morse synthesis) and the other must decode it with the built-in CW decoder. Drive both browsers via Playwright MCP, screenshot both sides, tear down. Bench-only (virtual rigs, no hardware, no real RF).

This skill is a sibling of packet-test / ft8-test and reuses the same bench (scripts/testrig.sh) and Playwright lifecycle. Read "Pitfalls" before debugging — CW shares FT8's RX-audio gate and needs the DECODE toggle on.

Prerequisites (verify before starting; fix and stop if missing)

  1. ./wfweb and ./tools/virtualrig/virtualrig are built. If not: qmake wfweb.pro && make -j$(nproc) and cd tools/virtualrig && qmake && make -j$(nproc). (virtualrig must include CW synthesis — synthesizeCw in tools/virtualrig/src/virtualrig.cpp; any recent build has it.)
  2. Repo .mcp.json includes --ignore-https-errors for the playwright MCP. Verify: pgrep -af @playwright/mcp must show --ignore-https-errors. If .mcp.json was just edited, the user must restart Claude Code (a /mcp reconnect re-uses cached args). Stop and tell them.
  3. Load the playwright tools via ToolSearch: select:mcp__playwright__browser_navigate,mcp__playwright__browser_snapshot,mcp__playwright__browser_click,mcp__playwright__browser_type,mcp__playwright__browser_take_screenshot,mcp__playwright__browser_wait_for,mcp__playwright__browser_evaluate,mcp__playwright__browser_close,mcp__playwright__browser_tabs,mcp__playwright__browser_select_option.

Boot the bench

./scripts/testrig.sh up 2

Both rigs come up on 14.074.000. The skill puts both into CW mode on the same frequency, so audio gates (matching freq+mode) and A's keyed morse reaches B as RX audio.

Per-tab UI bring-up (do for both A and B)

CW is a rig mode (unlike FT8's bar toggle): you switch the rig into CW, then open the CW bar.

  1. browser_navigate to the rig URL.
  2. Wipe stale browser state: browser_evaluate () => { localStorage.clear(); sessionStorage.clear(); }, then browser_navigate to the same URL again.
  3. Click the CLICK TO START splash (gesture needed for AudioContext).
  4. Wait for RX audio. The CW decoder taps the RX-audio playback chain, so audio must be flowing. The splash enables it ~1 s asynchronously — poll until window.audioEnabled === true before continuing.
  5. Switch to CW mode: window.setMode('CW'). This is async (round-trips to the rig) — poll until window.currentMode === 'CW' (a few hundred ms).
  6. Open the CW bar: window.renderCWOverlay() (equivalently the bottom-bar KEY button, which calls toggleCWBar()). Confirm #cwBar no longer has class hidden.
  7. Turn the decoder on: click #cwDecoderToggle (the DECODE button). It gains class active and textContent DECODE. Do this on both tabs (each side decodes the other). Give it ~1 s to spin up its worklet.

When refs go stale across snapshots, drive directly via browser_evaluate with stable IDs (the normal path, not a fallback):

  • Buttons: cwDecoderToggle (DECODE on/off), cwStopBtn, cwCloseBtn, cwSpeedUp, cwSpeedDown, cwLogBtn
  • Fields: cwInput (type-to-send), cwSpeedInput, cwCallInput, cwQsoInput
  • TX transcript (what THIS station is sending): #cwDisplay (and the cwCharacters global)
  • RX decode output (what this station COPIED): #cwDecoderTextInner .textContent ← assert on this
  • Functions (global): setMode('CW'), renderCWOverlay(), toggleCWBar(), sendCWText(text)
  • State (globals on window): currentMode, cwSpeed (default 20 wpm), audioEnabled

QSO sequence (single bidirectional run)

  1. A → key a string: on tab A, window.sendCWText('CQ TEST DE W1AAA'). This queues {cmd:'sendCW'}; virtualrig synthesizes the morse at the rig's CW pitch and routes it to B. At 20 wpm this 16-char string plays in ~8–10 s.
  2. Read B's copy: on tab B, poll document.getElementById('cwDecoderTextInner').textContent until it contains the sent text (allow ~30 s; give a few extra seconds for trailing characters). On the clean virtual channel at 20 wpm the copy is exact (CQ TEST DE W1AAA), but assert tolerantly — require the callsign W1AAA (and ideally TEST/CQ) to appear, not a byte-exact match, since a real decoder can drop or mangle a character.
  3. B → key the reply: on tab B, window.sendCWText('W1AAA DE W1BBB GE').
  4. Read A's copy: on tab A, poll #cwDecoderTextInner until it contains W1BBB.
  5. Screenshot both with browser_take_screenshot, fullPage: true. Save as .playwright-mcp/wfweb-cw-A.png and ...-B.png. (Bare names land in the project root; /tmp is rejected — the MCP roots are the project dir and .playwright-mcp/.)
Show full SKILL.md (305 more words)Show less

Teardown and report

  1. browser_close (both tabs), then ./scripts/testrig.sh down.
  2. Report PASS/FAIL and the two screenshot paths. PASS = each side copied the other's transmission (the expected callsign appears in its #cwDecoderTextInner).

Pitfalls — read before debugging

  • Two gates, not one. The decoder needs (a) window.audioEnabled === true (RX audio flowing — enabled ~1 s after CLICK TO START) and (b) the DECODE toggle (#cwDecoderToggle) clicked active. Miss either and #cwDecoderTextInner stays empty.
  • setMode('CW') is async. Poll window.currentMode === 'CW' before keying — keying while still in USB means no CW is generated and the receiver isn't gated to hear it.
  • Both rigs must be in CW on the same freq. Audio gates on matching freq+mode (same as packet). If one side is still USB, the mixer won't route the morse to it.
  • Don't confuse TX transcript with RX decode. #cwDisplay / cwCharacters is the sender's outgoing text (what it's keying). The received copy is #cwDecoderTextInner on the other tab. Assert on the receiver's #cwDecoderTextInner.
  • Assert tolerantly. The CW decoder is heuristic. On the clean virtual channel at 20 wpm it's a perfect copy, but write the assert as "contains the callsign", not byte-equality, so a single mis-decode doesn't fail a genuinely-working run. If you want maximum robustness, slow down (cwSpeed / #cwSpeedDown) before keying.
  • State is window globals, not a namespaced window.X.state object.
  • Don't skip localStorage.clear() — clean-slate mirrors a fresh user and avoids stale CW speed/mode.
  • Stale Playwright ref= handles re-render constantly; drive via browser_evaluate + the stable IDs above.
  • Screenshots are sandboxed to the project dir / .playwright-mcp/; /tmp is rejected.
  • cwLogBtn (LOG QSO) writes a scratch logbook, not the station log. testrig.sh starts each instance with --logbook .testrig/logbook_<i>.adi --no-remote-log; a wfweb started by hand without --logbook opens the operator's real logbook — never log a bench QSO against one.
  • Never key a physically-connected rig — this skill is virtual-bench only (no real RF).

© adecarolis, GPL-3.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .claude/skills/cw-test of adecarolis/wfweb.

Open the folder on GitHubat commit 7cd5d2f

Compare with similar skills

Cw Test next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Cw Test compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Cw Test this skilladecarolis/wfweb106—~1.9kAutomated safety check: PassGPL-3.0
Invisible Playwrightfeder-cr/invisible_dots32k—~989Automated safety check: PassMIT
Playwright E2E Testsonyx-dot-app/onyx32k1 repos~2.8kAutomated safety check: NotesCustom licence
Shogun Screenshotyohey-w/multi-agent-shogun1.4k—~901Automated safety check: NotesMIT
Playwright Debugvoicetreelab/voicetree923—~1.2kAutomated safety check: PassCustom licence
Hands On Testktnyt/cclsp675—~1.7kAutomated safety check: PassMIT

Similar skills

  • Invisible Playwright

    feder-cr/invisible_dots

    Use the Dot's browser (invisibleplaywright) for any task on a website: logging in, reading pages, filling forms, clicking through a site as a person would.

    32k GitHub stars~989 tokensUpdated today
    Testing & QAAuto-check passed
  • Playwright E2E Tests

    onyx-dot-app/onyx

    Write and maintain Playwright end-to-end tests for the Onyx application.

    32k GitHub starsUsed in 1 repo~2.8k tokens
    Testing & QAAuto-check: notes
  • Shogun Screenshot

    yohey-w/multi-agent-shogun

    スクリーンショットの取得・加工を行う。ローカルスクショから最新画像を取得、 PlaywrightでWebページをキャプチャ、画像のトリミング・リサイズ、機微情報を黒塗りマスキング。

    1.4k GitHub stars~901 tokensUpdated 2 mo ago
    Testing & QAAuto-check: notes
  • Playwright Debug

    voicetreelab/voicetree

    This skill should be used when the user asks to "debug the electron app", "connect playwright to VoiceTree", "take screenshots of the running app", "interact with the live UI", "inspect the running…

    923 GitHub stars~1.2k tokensUpdated today
    Testing & QAAuto-check passed
  • Hands On Test

    ktnyt/cclsp

    Performs manual hands-on testing of a web application using playwright-cli.

    675 GitHub stars~1.7k tokensUpdated 7 mo ago
    Testing & QAAuto-check passed
  • E2E Verification

    Chorus-AIDLC/Chorus

    A skill your agent uses when manually verifying a Chorus frontend change in a real browser — finding local login credentials, driving the running dev server with the Playwright MCP, logging in…

    1.2k GitHub stars~1.5k tokensUpdated yesterday
    Testing & QAAuto-check: notes

More from adecarolis/wfweb

  • Freedv Test

    adecarolis/wfweb

    Run an end-to-end FreeDV RADE EOO-callsign round-trip between two virtual wfweb rigs.

    106 GitHub stars~2.8k tokensUpdated today
    Auto-check passed
  • Ft8 Test

    adecarolis/wfweb

    Run an end-to-end FT8 decode round-trip between two virtual wfweb rigs.

    106 GitHub stars~2k tokensUpdated today
    Auto-check passed
  • Js8 Test

    adecarolis/wfweb

    Run an end-to-end JS8 message round-trip between two virtual wfweb rigs.

    106 GitHub stars~2.2k tokensUpdated today
    Auto-check passed
  • Packet Test

    adecarolis/wfweb

    Run an end-to-end AX.25 packet QSO test between two virtual wfweb rigs at 300 bd and 1200 bd.

    106 GitHub stars~1.7k tokensUpdated today
    Auto-check passed

Categories

Questions about Cw Test

What does Cw Test do?

Run an end-to-end CW (morse) decode round-trip between two virtual wfweb rigs. Cw Test is an agent skill from adecarolis/wfweb. Run an end-to-end CW (morse) decode round-trip between two virtual wfweb rigs.

When should I use Cw Test?

Cw Test fits situations like: the user asks to test CW; verify the CW decoder; validate the keyer TX / decode RX path.

How do I install Cw Test in Claude Code?

Run `npx skills add adecarolis/wfweb --skill cw-test -a claude-code`. Or copy the skill folder (.claude/skills/cw-test in adecarolis/wfweb) into .claude/skills/cw-test in your project. Claude Code loads it when a task matches its description.

How do I install Cw Test in Codex?

Run `npx skills add adecarolis/wfweb --skill cw-test -a codex`. Or copy the skill folder (.claude/skills/cw-test in adecarolis/wfweb) into .agents/skills/cw-test in your project. Codex loads it when a task matches its description.

Can I use Cw Test in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add adecarolis/wfweb --skill cw-test -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/cw-test, .gemini/skills/cw-test, .github/skills/cw-test and .opencode/skills/cw-test in your project.

What does Cw Test need to run?

Going by SKILL.md and its folder, Cw Test needs the command-line tools its instructions call (make).

Does Cw Test access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Cw Test safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Cw Test use?

Cw Test is published under the GPL-3.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Cw Test use?

About 1.9k tokens (SKILL.md is roughly 7.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Cw Test?

Skills that share tags, products or a category with Cw Test: Invisible Playwright (feder-cr/invisible_dots, 32k stars), Playwright E2E Tests (onyx-dot-app/onyx, 32k stars), Shogun Screenshot (yohey-w/multi-agent-shogun, 1.4k stars) and Playwright Debug (voicetreelab/voicetree, 923 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Cw Test?

adecarolis (a GitHub user) maintains it in adecarolis/wfweb, which has 106 GitHub stars. The repository holds 5 skills in this directory. The repository was last updated on October 7, 2026.

Source: adecarolis/wfweb on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.