feat(desktop): configurable group chat round cap + model-aware limits + token budget guard

[beplee]

Summary

• Replace hardcoded GROUP_CHAT_MAX_ROUNDS constant with config.yaml-driven group_chat.max_rounds (default 3, range 1-20)
• Add model-aware cap: free models (:free suffix or free_models allowlist) use group_chat.max_rounds_free (null = unlimited, clamped to HARD_CAP=20)
• Add token-budget guard: cumulative room-log tokens exceeding group_chat.token_budget (default 80k chars) force exitKind='capped' independent of round count — the real safety net for pathological cascades
• Add getGroupChatConfig(), isFreeModel(), getEffectiveRoundCap(), getCurrentModelName() helpers
• Add config.yaml schema in cli-config.yaml.example under group_chat:

Why

Hardcoded 3-round cap is too tight for free-model rooms (where API cost is zero) and too loose for paid-model rooms (where each round burns real money). A single config block drives everything automatically.

Config

group_chat:
  max_rounds: 3              # paid models
  max_rounds_free: null      # null = unlimited (HARD_CAP=20 backstop)
  token_budget: 80000        # hard ceiling per drive (chars, ~20k tokens)
  free_models:
    - upstage/solar-pro4
    - meituan/longcat-2.0
    - tencent/hy3
  hard_cap: 20               # absolute ceiling even for free models

Safety

• Token budget is the real backstop — fires regardless of model type or round count
• HARD_CAP=20 prevents unbounded free-model cascades (20 rounds × 6 bots = 120 messages max)
• Paid models always capped at max_rounds (default 3)
• exitKind='capped' reason logged for every termination path

Test plan

• Paid model → 3-round cap fires
• Free model (allowlist) → 20-round cap fires
• Token budget exhaustion → mid-drive cap fires
• Model switch mid-drive → cap re-evaluates
• 6-bot free room → halts cleanly at token budget
• Empty free_models + :free suffix → allowlist is sole gate (no bypass)

Refs: #96553

[alt-glitch]

This was generated by AI during triage.

Related to #92213: both extend Desktop Bot Mode limits, but #92213 persists per-room limits while this PR proposes global model-aware configuration and a token budget.

[beplee]

Closure rationale

This PR is closed as written because its target file () is deleted by #96726 (OutThisLife, "Rebuild Bot Mode on the app's design system", 230 files, 36k+/32k- diff). #96726 is moving toward merge and the file this PR patches will not exist on after that merge.

The feature itself is still needed. #96726 hardcodes at and the maintainer's comment on #96726 explicitly flags #96842 as the gap-filler:

Real contributor work that does NOT survive and needs re-implementation on the new modules: #96842 (round cap still hardcoded at , no token budget)

Round-cap constant — is newly hardcoded where an open contributor PR (#96842) makes it configurable; a config hook here (or a note ceding that to the follow-up) avoids re-burying the same constant.

What comes next: port the configurable round cap + model-aware limits + token-budget guard to the new module surface (, , ) after #96726 merges. The design work is already done in (branch ). Slot freed — we can open the port PR cleanly when #96726 lands and we have a slot.