Add Zoom Web co-host automation: auto-grant Multi-Pin on raised hand

[Copilot]

Tampermonkey userscript that runs under a Zoom Host/Co-Host session and automatically grants Multi-Pin permission to participants who raise their hand — no backend, no Zoom API, DOM-only.

Files

• scripts/zoom-host-tools.user.js — Main userscript. Matches *.zoom.us/wc/* and *.zoom.us/j/*.
• selectors/zoom-dom-selectors.json — Canonical selector map; update here when Zoom changes its UI, then mirror into the SELECTORS constant in the script.
• docs/automation-design.md — Architecture, DOM strategy, assumptions, extension points.
• docs/testing-checklist.md — Manual test plan across all three phases.

Phase 1 — Implemented (MVP)

Polls participant rows every CONFIG.SCAN_INTERVAL_MS (2.5 s):

1. Detects raised-hand indicator via layered fallback: data-testid → aria-label → class → text/emoji content
2. Derives stable participant ID (data-participant-id → display name → DOM fingerprint); deduplicates via Set
3. Hovers row to reveal hidden "More" button → opens menu → retries up to MENU_CLICK_RETRIES times for the Multi-Pin item
4. Guards against re-granting by checking for "Remove Multi-Pin" text in the open menu

// Selector resolution — iterates candidates, logs every match/miss
function resolveElement(candidates, root = document) {
  for (const selector of candidates) {
    const el = root.querySelector(selector);
    if (el) return el;
  }
  return null; // caller handles null safely
}

Phases 2 & 3 — Scaffolded

• Phase 2: checkCameraStatus() fires after every grant; reads video icon aria-label/class to infer on/off state. TODO (Phase 2) marks where to wire in the camera-off chat reminder.
• Phase 3: MutationObserver on the chat container flags messages matching http(s)://, t.me/, bit.ly/, discord.gg/. TODO (Phase 3) marks the moderation hook.

Selector strategy

All selectors live in selectors/zoom-dom-selectors.json with ordered candidates arrays and fallbackStrategy prose per element. Priority: data-testid → aria-label → class name → text match → structural traversal. Selector mismatches emit [WARN] logs; successful matches emit [DEBUG] logs — keeping breakage easy to diagnose.

Debug panel

When CONFIG.DEBUG = true, a floating overlay shows live scan/grant counters and last action taken.

Original prompt

Here’s the Copilot Super-Agent Query version — stronger, more explicit, and designed so the agent builds a more reliable Zoom automation foundation while still keeping the first delivery small.

COPILOT SUPER-AGENT QUERY — ZOOM CO-HOST MULTIPIN AUTOMATION

You are a coding agent working inside this repository.

Your task is to build a lightweight Zoom Web automation tool that runs under a Zoom Host or Co-Host account and automates Multi-Pin granting for participants who raise their hand.

The main goal is to create a simple first working version quickly, but structure it so it can later support camera checks, chat moderation, and other host/co-host workflows.

Do not overengineer.
Do not build a backend.
Do not require external services.
Prefer plain JavaScript and a Tampermonkey userscript that interacts directly with the Zoom Web DOM.

Assume the automation user already has the required co-host or host permissions to perform these actions.

⸻

PRIMARY OBJECTIVE

Implement this core behavior:

When a participant raises their hand, automatically:
1. detect the participant row
2. determine whether that participant already has Multi-Pin permission
3. if not, open the participant menu
4. click Allow to Multi-Pin
5. store internal state so the same participant is not processed repeatedly

This is the most important feature.
Everything else is secondary.

⸻

PERMISSION MODEL

Assume the script runs in a Zoom session where the user has:
• Host permissions, or
• Co-Host permissions

So the script may safely attempt actions such as:
• opening participant action menus
• granting Multi-Pin
• sending chat messages
• performing moderation actions later

Do not implement authentication logic.
Do not use Zoom APIs unless absolutely necessary.
The tool should work through the existing Zoom Web interface.

⸻

BUILD STRATEGY

Implement the project in phases, but code the repository so Phase 1 works immediately.

Phase 1 — Required

Auto-grant Multi-Pin when a participant raises their hand.

Phase 2 — Scaffold only

After granting Multi-Pin, detect whether the participant’s camera is on.
If the camera is off, prepare logic to send a one-time chat message:

Please turn your camera on to use Multi-Pin.

This phase can be partially scaffolded if camera detection is uncertain.

Phase 3 — Scaffold only

Monitor chat for likely spam links:
• http://
• https://
• t.me
• bit.ly
• discord.gg

For now, just log detections and structure future moderation hooks.

⸻

CRITICAL REQUIREMENT: DOM MAPPING LAYER

Because Zoom Web UI may change, you must create a DOM selector mapping layer.

Do not hardcode selectors everywhere.

Create a centralized selector/config system that contains entries for:
• participant list container
• participant row
• raised hand indicator
• participant name
• participant menu button
• Multi-Pin menu item
• camera status indicator
• chat container
• chat message rows
• chat input

The code should use helper functions to resolve selectors from this config.

If the exact selector is uncertain, implement a fallback strategy:
• first try stable attributes
• then aria-labels
• then text matching
• then defensive DOM traversal

Comment all assumptions clearly.

⸻

RELIABILITY REQUIREMENT

The agent must make the automation as reliable as possible without making it heavy.

Implement:
• a lightweight polling loop or MutationObserver
• max scan interval of about 2–3 seconds
• deduplication of already-processed participants
• guard clauses to prevent repeated clicks
• safe retry behavior when menus fail to open
• console logging for every important decision

Examples of logs:
• detected raised hand
• participant already has multipin
• granting multipin
• failed to find menu item
• selector mismatch warning

⸻

IMPLEMENTATION REQUIREMENTS

Use:
• plain JavaScript
• Tampermonkey userscript format
• no backend
• no external libraries
• no framework unless absolutely necessary

Keep code modular with functions or modules for:
• selector resolution
• participant scanning
• raised-hand detection
• multipin status detection
• multipin granting
• camera status checks
• chat monitoring
• moderation hooks
• debug logging

⸻

FILES TO CREATE

Create this structure:

/scripts/zoom-host-tools.user.js
Main Tampermonkey userscript

/selectors/zoom-dom-selectors.json
Central selector map and fallback notes

/docs/automation-design.md
Explain architecture, assumptions, Zoom DOM strategy, and extension points

/docs/testing-checklist.md
Manual test plan for Zoom Web behavior

⸻

REQUIRED BEHAVIOR DETAILS

Raised Hand Detection

Detect participants who currently show the raised-hand state.

Multi-Pin Detection

Before clicking anything, try to determine whether the participant already has Multi-Pin enabled.

If direct status detection is not available, use one of these strategies:
1. infer from menu text
2. inspect whether the action says “Allow to Multi-Pin” vs some granted state
3....

---

📍 Connect Copilot coding agent with Jira, Azure Boards or Linear to delegate work to Copilot in one click without leaving your project management tool.

[cr-gpt]

Seems you are using me but didn't get OPENAI_API_KEY seted in Variables/Secrets for this repo. you could follow readme for more information

[cr-gpt]

Seems you are using me but didn't get OPENAI_API_KEY seted in Variables/Secrets for this repo. you could follow readme for more information

[FriskyDevelopments]

Tampermonkey userscript that runs under a Zoom Host/Co-Host session and automatically grants Multi-Pin permission to participants who raise their hand — no backend, no Zoom API, DOM-only.

Files

scripts/zoom-host-tools.user.js — Main userscript. Matches .zoom.us/wc/ and .zoom.us/j/.
selectors/zoom-dom-selectors.json — Canonical selector map; update here when Zoom changes its UI, then mirror into the SELECTORS constant in the script.
docs/automation-design.md — Architecture, DOM strategy, assumptions, extension points.
docs/testing-checklist.md — Manual test plan across all three phases.
Phase 1 — Implemented (MVP)

Polls participant rows every CONFIG.SCAN_INTERVAL_MS (2.5 s):

Detects raised-hand indicator via layered fallback: data-testid → aria-label → class → text/emoji content
Derives stable participant ID (data-participant-id → display name → DOM fingerprint); deduplicates via Set
Hovers row to reveal hidden "More" button → opens menu → retries up to MENU_CLICK_RETRIES times for the Multi-Pin item
Guards against re-granting by checking for "Remove Multi-Pin" text in the open menu
// Selector resolution — iterates candidates, logs every match/miss
function resolveElement(candidates, root = document) {
for (const selector of candidates) {
const el = root.querySelector(selector);
if (el) return el;
}
return null; // caller handles null safely
}
Phases 2 & 3 — Scaffolded

Phase 2: checkCameraStatus() fires after every grant; reads video icon aria-label/class to infer on/off state. TODO (Phase 2) marks where to wire in the camera-off chat reminder.
Phase 3: MutationObserver on the chat container flags messages matching http(s)://, t.me/, bit.ly/, discord.gg/. TODO (Phase 3) marks the moderation hook.
Selector strategy

All selectors live in selectors/zoom-dom-selectors.json with ordered candidates arrays and fallbackStrategy prose per element. Priority: data-testid → aria-label → class name → text match → structural traversal. Selector mismatches emit [WARN] logs; successful matches emit [DEBUG] logs — keeping breakage easy to diagnose.

Debug panel

When CONFIG.DEBUG = true, a floating overlay shows live scan/grant counters and last action taken.