fix(api): scope receipt access to owning tenant; RLS remediation#177
Open
chrismaz11 wants to merge 2 commits into
Open
fix(api): scope receipt access to owning tenant; RLS remediation#177chrismaz11 wants to merge 2 commits into
chrismaz11 wants to merge 2 commits into
Conversation
Enforce tenant ownership on receipt routes so an API key can only access its own tenant's receipts. Previously these routes resolved a receipt by id alone, exposing any receipt (and the large pool of legacy null-owner rows) to any authenticated key. Add `callerOwnsReceipt(authUserId, ownerUserId)` — a pure decision that matches a caller's resolved tenant identity (authContext.userId, from api_keys.user_id) against Receipt.userId, treating null/undefined on either side as no-match — and `loadReceiptOwner()`, a raw-SQL owner lookup (Receipt.userId is not on the Prisma model). Gate these routes; both not-found and not-owned return 404 so existence is not leaked across tenants: - GET /api/v1/receipt/:id - GET /api/v1/receipt/:id/pdf - POST /api/v1/receipt/:id/verify - POST /api/v1/anchor/:id Rewrite GET /api/v1/receipts to list only the caller's own receipts (empty list for a null identity). POST /receipt/:id/revoke is unchanged: it already requires an issuer signature under its own trust model. Add receiptAccess.test.ts: 7 unit tests over callerOwnsReceipt covering own-read, cross-tenant, null-owner, null/undefined/empty-string callers. Docs: correct the CLAUDE.md hashing note — receipt hash and inputs commitment are keccak256 (ethers), not SHA-256; SHA-256 is used only for zkp/zkml digests and ATTOM normalization. This is the application-code half of the RLS remediation; the backend-only RLS policies on the 10 previously RLS-disabled tables (tenants + 9 dormant) were applied separately via Supabase migrations. Co-Authored-By: Claude Opus 4.8 <[email protected]>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
Replace additive bucket accumulation with probabilistic OR scoring (1 − Π(1 − wᵢ)) so multiple weak signals accumulate without a single moderate signal saturating the score. Extracts the scoring logic into scoring.ts and refactors types/forensics/layout/patterns to carry per-signal weights. Adds documentExtraction.ts in apps/api to parse PDF buffers (text, geometry, structural tokens) into the ExtractedDocument the core RiskEngine consumes. Updates CLAUDE.md with accurate stub/prod caveats for risk, ZK, and ZKML modules. Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Application-code half of the RLS remediation workstream on the 10 previously RLS-disabled tables. Enforces tenant ownership on receipt routes and corrects a docs note. Backend-only RLS policies were applied separately via Supabase migrations (not in this diff).
Cross-tenant read-exposure fix
Previously these routes resolved a receipt by id alone, exposing any tenant's receipt (and the large pool of legacy null-owner rows) to any authenticated key. Now gated by
callerOwnsReceipt(authUserId, ownerUserId)(matchesauthContext.userId←api_keys.user_idagainstReceipt.userId; null/undefined either side = no match), with owner loaded via raw SQL sinceReceipt.userIdisn't on the Prisma model. Not-found and not-owned both return 404 so existence isn't leaked across tenants:GET /api/v1/receipt/:idGET /api/v1/receipt/:id/pdfPOST /api/v1/receipt/:id/verifyPOST /api/v1/anchor/:idGET /api/v1/receiptsrewritten to list only the caller's own receipts (empty list for a null identity).POST /receipt/:id/revokeis intentionally unchanged — it already requires an issuer signature under its own trust model.Tests
apps/api/src/receiptAccess.test.ts— 7 unit tests overcallerOwnsReceipt: own-read, cross-tenant, null-owner, null/undefined/empty-string callers.Docs
Corrected the CLAUDE.md hashing note: the receipt hash and inputs commitment are keccak256 (ethers), not SHA-256; SHA-256 is used only for zkp/zkml digests and ATTOM normalization.
Test plan
npm -w apps/api exec -- vitest run src/receiptAccess.test.ts(7 pass)🤖 Generated with Claude Code