feat: enforce free-tier node ceiling and MCP gate

[cameronapak]

Summary

Free outlines are now capped at 2,000 live nodes and MCP agent access is a paid capability, so the free/paid line is enforced server-side rather than by convention.

Fixes #170

Changes

1. A free user who would push past 2,000 nodes gets a "limit reached" upgrade toast and the insert rolls back; edits, moves, deletes, and net-reducing batches always apply, and an over-cap (downgraded) outline is never locked.
2. A free MCP token can still handshake and list tools, but every tools/call is refused with an in-protocol upgrade error — no read or write reaches the outline.
3. The DO gates writes at commit time (applyBatchGated/upsertNodesGated): it counts net node growth by distinct id inside the write transaction and refuses only batches that grow the outline past the cap it's told.
4. The refusal decision is a pure, exhaustively unit-tested batchExceedsNodeLimit; the Worker resolves the caller's plan once per node-creating write (skipping pure-delete batches) and maps a refusal to 403 {error:"node_limit",limit}.
5. An operator comps a user by inserting an active subscription row — getPlan already treats it as paid, so no Stripe round-trip and no new override path.
6. Review hardening: the plan lookup on the node path uses the billing user.id (not the DO-routing id, which collapses the owner to default), and a pure countNetGrowth closes a delete+reinsert count bypass.

Start here: change 6 — the owner billing-id fix and the delete+reinsert bypass, both found by adversarial review; the growth count now takes each id's last op.

Flow

flowchart LR
  A[structural insert] --> B[Worker: getPlan]
  B --> C[DO applyBatchGated]
  C -->|within cap| D[commit + echo]
  C -->|would exceed| E[403 node_limit]
  E --> F[toast + rollback]

Loading

Breaking / Migration

None. Additive: no schema change, no new Node field. Existing alpha accounts (no subscription row) are free and unaffected until they cross the cap.

Test plan

• bun run typecheck / typecheck:worker / typecheck:test — green
• bun test — passing, incl. batchExceedsNodeLimit boundary cases (under/at/over cap, downgraded outline never locked, net-reducing batch allowed), countNetGrowth (the delete+reinsert bypass case), and the MCP free-token gate (tools/call refused, discovery open)
• bun run lint, bun run fmt:check — clean
• Adversarial review: two Opus reviewers (correctness + security/paywall). No bypass; gate fails closed. Two findings fixed (owner billing-id on the node path; delete+reinsert cap-count bypass). Every node-creating path (batch, legacy upsert, PATCH, kv-as-nodes, ungated MCP write) traced and closed.
• Needs manual check: the live cap toast + optimistic rollback in a signed-in free session, and an MCP client hitting the refusal — needs a running Worker + D1 + a free account; e2e can't reach it (seedOutline mocks the Worker, so the real getPlan/DO cap never runs).

[coderabbitai]

Warning

Review limit reached

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

Next review available in: 37 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

Run ID: c3abf7fd-2ef3-4a83-86e2-26b42644ae92

📥 Commits

Reviewing files that changed from the base of the PR and between 9d6c1aa and ec51859.

📒 Files selected for processing (9)

• .changeset/free-tier-entitlements.md
• src/data/nodes-client-effect.ts
• src/data/structural.ts
• worker/index.ts
• worker/mcp.test.ts
• worker/mcp.ts
• worker/outline-do.ts
• worker/plan.test.ts
• worker/plan.ts

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/wayfinder-151-next-5b0145

---

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.}