feat: add MicProlongedUsage detection event

[devin-ai-integration]

feat: add MicProlongedUsage detection event

Summary

Adds a new DetectEvent::MicProlongedUsage variant that fires when an app has been using the mic for 3 minutes continuously. The frontend then decides whether to show a notification based on the current listener state and user settings. This enables prompting users to start recording when they appear to be in a meeting.

How it works:

Rust (detect plugin):

• When MicStarted is received for an app (that passes policy + cooldown filtering), a per-app timer is spawned with a 3-minute threshold using CancellationToken
• When the timer fires, it acquires the std::sync::Mutex state lock, calls claim() with the generation number to atomically verify and remove the entry, then emits MicProlongedUsage
• When MicStopped is received, the timer's CancellationToken is cancelled and the entry removed
• A 1-hour cooldown per app prevents repeated notifications for the same app. Expired cooldown entries are lazily cleaned up in is_in_cooldown() to prevent unbounded map growth.
• The tauri-plugin-listener dependency is removed from the detect plugin — the "is listening" check is intentionally moved to the consumer side (frontend)

Frontend:

• listener.tsx handles the new micProlongedUsage event, checking both the notification_in_meeting_reminder setting and store.getState().live.status === "active" before showing a notification
• New MeetingReminderToggle in Settings > Lab lets users enable/disable the feature (default: enabled)
• Notification: "Zoom has been using the mic for 3 min. Start listening?"

Race condition mitigations:

• Generation-based claiming: Each timer gets a monotonically increasing generation number. claim() only succeeds if the generation matches, preventing stale timers from emitting after a stop+restart cycle.
• CancellationToken: tokio::select! between sleep and cancellation ensures prompt abort on MicStopped.
• Cooldown: After a successful claim(), the app enters a 1-hour cooldown, suppressing duplicate notifications.

Key files changed:

• plugins/detect/src/mic_usage_tracker.rs — new module: MicUsageTracker with generation-based timers, cooldowns (with lazy expiry cleanup), and comprehensive unit tests
• plugins/detect/src/env.rs — new module: Env trait abstracting event emission + DnD checks for testability
• plugins/detect/src/handler.rs — refactored to use Env trait; added integration-style tests
• plugins/detect/src/events.rs — new MicProlongedUsage variant (replaces old From impl)
• plugins/detect/src/policy.rs — removed is_listening/skip_when_listening; added should_track_app(); user_ignored_bundle_ids changed to HashSet
• plugins/detect/src/lib.rs — split into DetectorState + ProcessorState; std::sync::Mutex instead of tokio::sync::Mutex
• apps/desktop/src/contexts/listener.tsx — handles micProlongedUsage with listener state + setting checks
• apps/desktop/src/components/settings/lab/index.tsx — new MeetingReminderToggle component

Review & Testing Checklist for Human

• TS bindings consistency — bindings.gen.ts was hand-edited. Regenerate via cargo test -p tauri-plugin-detect to verify it matches.
• Listener state check timing — Frontend checks store.getState().live.status === "active" at notification time. If user starts listening between timer fire and notification display, the notification is correctly suppressed. Verify this works as expected.
• DnD not checked for MicProlongedUsage — The timer fires regardless of DnD state (only MicStarted/MicStopped notifications respect DnD). Confirm this is intentional behavior.
• Manual test plan:
  1. Start a mic-using app (e.g. Zoom), wait 3+ minutes without Hyprnote listening → notification should appear
  2. Stop the mic app before 3 min → no notification
  3. Start Hyprnote listening before 3 min → no notification (frontend suppresses)
  4. Stop mic app right at ~3 min → no false notification (cancellation race)
  5. Stop then immediately restart mic app → new timer starts fresh
  6. After notification fires, stop and restart mic app → no notification for 1 hour (cooldown)
  7. Toggle off "In-Meeting Reminder" in Settings > Lab → no notifications

Notes

• Threshold is hardcoded at 3 minutes (MIC_ACTIVE_THRESHOLD). Cooldown is 1 hour (COOLDOWN_DURATION). Could be made configurable later.
• Timers are per-app: e.g. if Zoom and Slack are both using the mic, each gets its own independent timer.
• Comprehensive unit tests added for MicUsageTracker and integration-style tests in handler.rs.
• Cooldown entries are lazily cleaned up when is_in_cooldown() is called (not via periodic task), which is sufficient since the map is bounded by the number of distinct apps that have triggered timers.

Link to Devin run: https://app.devin.ai/sessions/0abeb320269b434cade8e4591a698c6e
Requested by: @yujonglee

[netlify]

✅ Deploy Preview for hyprnote-storybook canceled.

Name	Link
🔨 Latest commit	cde5528
🔍 Latest deploy log	https://app.netlify.com/projects/hyprnote-storybook/deploys/6985aaec8b55910008d79e46

[netlify]

✅ Deploy Preview for hyprnote canceled.

Name	Link
🔨 Latest commit	cde5528
🔍 Latest deploy log	https://app.netlify.com/projects/hyprnote/deploys/6985aaec9a106200081ec82f

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR that start with 'DevinAI' or '@devin'.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 4 additional findings.

[devin-ai-integration]

Devin Review found 1 new potential issue.

View 8 additional findings in Devin Review.

[devin-ai-integration]

Devin Review found 1 new potential issue.

View 9 additional findings in Devin Review.

[devin-ai-integration]

Devin Review found 1 new potential issue.

View 10 additional findings in Devin Review.