1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
//! Token Maxing Mode (RFC-031).
//!
//! Autonomous burn of subscription-quota providers during a user-configured
//! time window. The hard rule: **only subscription providers** are eligible —
//! metered keys can never be silently drafted into a maxing run.
//!
//! Architecture:
//!
//! - [`config`] — TOML schema (`[token-maxing]`).
//! - [`budget`] — `ProviderBudget`, the self-tracked counter. Reuses the
//! `BudgetManager` window+reset pattern, re-keyed by provider.
//! - [`quota_tracker`] — `QuotaTracker`, the decision layer that merges
//! self-tracked, recalibration, and reactive 429 signals into one
//! `Availability` verdict per provider.
//! - [`planner`] — `WorkPlanner`, three-source task synthesis (autonomous
//! skills → projects/mounts → recurring patterns). (Phase 3.)
//! - [`maxer`] — `TokenMaxer`, the drain → rotate → wait → resume loop.
//! (Phase 3.)
//! - [`session`] — `TokenMaxingSession` + persisted report. (Phase 3.)
//!
//! The non-obvious invariant is the eligibility check: providers missing
//! from `[token-maxing.providers]` (or any with `billing_model !=
//! "subscription"`) are **never** eligible. This is the single choke point
//! that upholds the user's "절대 동작하면 안 된다" constraint.
pub use ;
pub use ;
pub use TokenMaxer;
pub use ;
pub use ;
pub use ;