jackin-usage

Usage, pricing, telemetry, and token monitors for the jackin-capsule daemon. Owns the in-capsule accounting of agent token spend/cost and the telemetry store that persists it.

What this crate owns

Token monitoring (token_monitor) and usage accounting (usage) for running agents.
Telemetry emission + store (telemetry, telemetry_store, logging) — the capsule-side observability tier.
Usage output shaping (output). (clog!/cdebug! logging infrastructure is re-exported from jackin-diagnostics.)

Architecture tier and allowed dependencies

Infrastructure (capsule-side observability/accounting). Allowed inward dependencies: jackin-core, jackin-protocol, jackin-diagnostics. No dependency on jackin-capsule (would be circular), jackin-tui, jackin-console, or any presentation crate. Logging infrastructure (logging, clog!, cdebug!) lives here so both binaries share one tier.

Structure

Module	Owns	Tests
`lib.rs`	crate root, re-exports	—
`token_monitor.rs` · `token_monitor/`	token spend monitoring	`tests.rs`
`usage.rs` · `usage/`	usage/pricing accounting	`tests.rs`
`telemetry.rs`	telemetry emission	—
`telemetry_store.rs` · `telemetry_store/`	persistent telemetry store	`tests.rs`
`store_backend.rs`	turso SQLite import chokepoint	—
`logging.rs` · `logging/`	shared logging tier (`clog!`/`cdebug!`)	`tests.rs`
`output.rs`	usage output shaping	—

Public API

Token-monitor + usage-accounting types consumed by jackin-capsule; the shared clog!/cdebug! logging tier re-exported across the workspace. Avoid cloning full usage views during account materialization — serialize from borrowed views/iterators (tracked perf item).

How to verify

cargo nextest run -p jackin-usage
cargo clippy -p jackin-usage --all-targets -- -D warnings