Desktop productization: always-on lifecycle and high-frequency workflow bundles

[webup]

Migrated from issue #117, originally opened by @Thalia325. Moved to a discussion for broader design conversation.

Related capability

Applications (能应用)

Problem or motivation

ClawMaster already has a complete admin console skeleton and a working desktop shell, but two gaps stand out from a production-grade product and enterprise-endpoint perspective:

1. Desktop lifecycle capabilities are not yet productized. The Settings page already exposes desktop options — launch at login, show tray icon, minimize to tray — but they are still disabled placeholders, and the README marks the desktop app as beta. That makes today's desktop closer to "a runnable desktop admin console" than to "a long-running, low-disturbance, always-on desktop companion." For individual users and enterprise knowledge workers, this meaningfully weakens daily stickiness.

2. Scene-oriented workflow bundles are still sparse. Existing scene bundles focus on photo-qa, invoice, and coding-boost. That shows the system can package focused capabilities, but relative to the "real work / real life" positioning it still lacks high-frequency, low-friction user-facing workflows — meeting notes, email/IM summaries, personal knowledge capture, household/admin document organization, etc.

For a production-grade product, a raw capability catalog is not enough. Without a set of ready-to-use high-frequency workflows, typical users hit: they know the system is powerful but don't know where to start, they know OCR / memory / skills exist but can't compose them into actual flows, and they struggle to form a daily habit.

Proposed solution

Treat "desktop productization" and "high-frequency scene bundles" as one unified product hardening initiative.

Phase 1 — Land always-on desktop lifecycle

1. Launch at login — per-platform registration, persisted preference, explicit messaging on unsupported platforms.
2. Tray icon — persistent tray presence, basic tray menu, restore-main-window / quit / view-status entries.
3. Minimize to tray — continue running in background after minimize; offer "quit" vs "stay in tray" on window close.
4. Lifecycle consistency — avoid invisible zombie processes, avoid "user expects background run but app actually quit," define graceful-shutdown behavior and background-task teardown.

Phase 2 — Ship high-frequency workflow bundles
Prioritize these scenes:

1. Meeting notes — import transcript/docs, auto-summarize, generate action items, persist into memory for future recall.
2. Email / IM summary — aggregate messages, produce a highlights digest, output reply drafts or a todo list.
3. Knowledge capture — collect source material, structure it, persist into memory / knowledge workflow, retrievable and reusable later.
4. Document / receipt archiving — OCR parse, structured field extraction, archive for later search.

Phase 3 — Connect desktop entry points to workflows

1. Desktop surfaces quick-enter affordances for common scenes.
2. Users trigger common tasks cheaply from the always-on state.
3. The desktop evolves from "a shell" into "a high-frequency work entry point."

Prefer reusing existing capabilities instead of rebuilding: PaddleOCR Doc Parsing, Content Draft, memory-clawmaster-powermem, Find Skills, and the existing setup / skills / memory / OCR pages.

Alternatives considered

1. Keep expanding only the skill catalog — helps power users, but doesn't improve the onboarding path for regular users.
2. Leave desktop settings as placeholders and defer to a later release — lets the UI stay permanently ahead of real capability and erodes trust in product maturity.
3. Build each workflow from scratch as a brand-new capability — higher cost, less efficient than reusing existing OCR, memory, draft, and skill chains.