[2witstudios]

Summary

Follow-up from the post-merge audit of #2258 (billing slice, issue #2260). One sandbox was billing through two different payer-resolution rules depending on charge type — runtime billing resolved from the caller's client-influenceable surface (ctx.tenantId/ctx.agentPageId), storage billing already resolved from the session's own driveId/ownerId. The M5 fix (session un-conflation) made those diverge for real: one session can host conversations from a different drive than the one it was spawned in.

• Unify payer resolution on the session (withMachineBilling in tool-runners.ts): added a cheap resolveBillingSession seam (session-row read, no provisioning) so the payer resolves from the ACQUIRED session's driveId/ownerId before gating — preserving the "credit-exhausted never wakes the sandbox" property. SandboxBillingDeps.resolvePayerId now takes {driveId, ownerId} instead of {tenantId, agentPageId}.

• One rule, written once: added resolveSessionPayerId (packages/lib/src/billing/sandbox-payer.ts) — drive owner else session owner, the same rule storageBillingTarget already applied for storage. Deleted resolveSandboxPayerId, lookupPageOwnerId, and the dead resolveAgentSessionPayerId (the wrong-shaped page-keyed twin), and fixed the now-false "resolves correctly BY CONSTRUCTION" comments this invalidated across sandbox-billing.ts, sandbox-storage-attribution.ts, sandbox-storage-reconcile.ts, and usage-breakdown-query.ts.

• no_session added to AUTHZ_DENY_REASONS so the expected "conversation predates sessions" refusal logs at warn, not error.

• Reconcile totals accounting fix: a chargeStorage success followed by a throwing advanceAgentSessionWatermark now counts toward charged/totalCostDollars (the money DID move) with a new chargedButUnadvanced counter, instead of silently under-reporting it as failed.

• First-class drive/session attribution on trackUsage (Terminal Epic 3 usage-breakdown fix): added a sessionId field alongside the existing driveId on AIUsageData/writeAiUsage/ai_usage_logs. sandbox-storage-billing.ts's chargeStorage now passes driveId/sessionId top-level (previously JSON-forensics-only in metadata). The realtime shell-handler's three trackUsage call sites now pass the session's real driveId/agentSessionId instead of a permanently-undefined pageId.

  Note for fix(realtime): hoist cross-tier tenant/membership resolution into lib; dead-code sweep; retire machine vocabulary in live identifiers [MED] #2265: the new sessionId field on AIUsageData/writeAiUsage is the seam to consume for realtime shell-usage attribution — it's self-contained in packages/lib (schema column already existed, just unused).

• Corrected stale conversation≡session comments (tool-runners.ts, sandbox-tools-runtime.ts, shell-handler.ts's ShellSandboxResult) and test-hygiene copy-paste leftovers.

Key acceptance test (tool-runners.test.ts): a session whose driveId differs from the caller's surface drive bills the SESSION's payer for the runtime stream — fails on master's withMachineBilling, passes after this fix. The storage stream already had equivalent coverage in sandbox-storage-reconcile.test.ts.

Test plan

• bun run typecheck (root) — 16/16 tasks pass
• bun run lint (root) — 14/14 tasks pass
• cd packages/lib && bun run typecheck — clean
• bun run knip:check — 4 issues, all within baseline (no new dead exports)
• bun run test:unit — lib: 8632 passed / 11 skipped, 0 failed. web: 15057 passed, 1 failed (grouping.test.ts midnight/TZ test — pre-existing, documented env-only flake, unrelated file untouched by this diff)
• bun run test:security — 44/50 suites pass; the 6 failures (Session Service, Device Auth Utilities, Permissions, Login Route, Signup Route, Mobile Login) are pre-existing stale suites unrelated to this change (deleted routes / excluded configs)
• New/updated tests: packages/lib/src/billing/__tests__/sandbox-payer.test.ts, packages/lib/src/services/sandbox/__tests__/{tool-runners,sandbox-billing,sandbox-storage-billing,sandbox-storage-reconcile}.test.ts, apps/web/src/lib/ai/tools/__tests__/sandbox-tools-runtime.test.ts, apps/realtime/src/terminal/__tests__/shell-handler.test.ts, apps/web/src/app/api/cron/reconcile-machine-storage/__tests__/route.test.ts

Closes #2260

🤖 Generated with Claude Code

https://claude.ai/code/session_01ExHmbFs9zbegha2bQDuL2b

[coderabbitai]

Warning

Review limit reached

@2witstudios, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 23 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 6df0a566-3742-46a7-a1a6-e7b676f4498c

📥 Commits

Reviewing files that changed from the base of the PR and between 4720452 and 3abaf6b.

📒 Files selected for processing (21)

• apps/realtime/src/terminal/__tests__/shell-handler.test.ts
• apps/realtime/src/terminal/shell-handler.ts
• apps/realtime/src/terminal/terminal-session-map.ts
• apps/web/src/app/api/cron/reconcile-machine-storage/__tests__/route.test.ts
• apps/web/src/app/api/cron/reconcile-machine-storage/route.ts
• apps/web/src/lib/ai/tools/__tests__/sandbox-tools-runtime.test.ts
• apps/web/src/lib/ai/tools/sandbox-tools-runtime.ts
• apps/web/src/lib/subscription/usage-breakdown-query.ts
• packages/lib/src/billing/__tests__/sandbox-payer.test.ts
• packages/lib/src/billing/sandbox-payer.ts
• packages/lib/src/logging/logger-database.ts
• packages/lib/src/monitoring/ai-monitoring.ts
• packages/lib/src/services/sandbox/__tests__/sandbox-billing.test.ts
• packages/lib/src/services/sandbox/__tests__/sandbox-storage-billing.test.ts
• packages/lib/src/services/sandbox/__tests__/sandbox-storage-reconcile.test.ts
• packages/lib/src/services/sandbox/__tests__/tool-runners.test.ts
• packages/lib/src/services/sandbox/sandbox-billing.ts
• packages/lib/src/services/sandbox/sandbox-storage-attribution.ts
• packages/lib/src/services/sandbox/sandbox-storage-billing.ts
• packages/lib/src/services/sandbox/sandbox-storage-reconcile.ts
• packages/lib/src/services/sandbox/tool-runners.ts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}