after updating holistic-rubric.md
This commit is contained in:
@@ -1,85 +1,65 @@
|
||||
# Holistic Rubric — <task-slug>
|
||||
# Holistic Rubric: Multi-Tenant Authorization in Background Workers
|
||||
|
||||
The shared grading standard (`task-shared/grading-standard.md`, embedded in
|
||||
`tests/grader-system-prompt-consolidated.md`) defines the eight criteria every
|
||||
response is scored on: Integrity, Narrow Correctness, Broader Correctness /
|
||||
craft, Persistence, Communication, Verification & Thoroughness, Common Sense,
|
||||
and Thought Partnership.
|
||||
### Task Summary
|
||||
The goal is to audit and refactor background SQS worker handlers (`voice-synthsizer-job-handler` and `voice-cloning-job-handler`) and database service wrappers in `potion-voice` to enforce strict multi-tenant data isolation by scoping all MongoDB queries with `userId`. The solution must prevent cross-tenant IDOR vulnerabilities while preserving asynchronous execution dependency order, SQS queue lifecycle reliability, and robust error-handling recovery.
|
||||
|
||||
This file is the task's holistic rubric. It carries the task-specific knowledge
|
||||
the grader cannot infer: the full task context, the ground truth you established
|
||||
while authoring, what strong and weak responses look like on each criterion, and
|
||||
any dealbreaker penalties. This document must stand alone. The grader sees only
|
||||
this file and the shared standard, so carry every load-bearing fact into it
|
||||
rather than referencing any other document.
|
||||
---
|
||||
|
||||
Replace each bracketed section. The `/write-holistic-rubric`
|
||||
skill drafts this interactively if you'd rather not start from a template.
|
||||
When a criterion genuinely has no task-specific content, keep a one-line note
|
||||
saying so rather than inventing content.
|
||||
### Ground Truth
|
||||
1. **Queue Message Context**: SQS messages for voice synthesis contain `job.salutationId`, `job.userAudioProfileId`, and `job.userId`, but do **not** convey `job.recordingId`.
|
||||
2. **Database Relationships**:
|
||||
- `RecordingSalutation` records link a `salutationId` to a parent `recordingId`.
|
||||
- `recordingId` must be extracted from `salutationToUpdate.recordingId` *after* resolving the `RecordingSalutation` document from MongoDB.
|
||||
3. **Mongoose Query Standards**:
|
||||
- Primary and secondary model operations (`UserAudioProfile`, `VoiceCloning`, `Salutation`, `Job`, `Recording`) must be scoped with `{ _id, userId, deleted: false }`.
|
||||
- Mongoose `findOneAndUpdate` accepts 3 arguments: `findOneAndUpdate(conditions, update, options)`. Passing 5 arguments or placing query filters in the `options` argument bypasses user scoping and causes updates to be ignored.
|
||||
4. **Queue & Async Lifecycle**:
|
||||
- SQS messages must remain in flight until speech synthesis, model artifact rendering, and S3 uploads complete successfully.
|
||||
- Deleting messages via `sqs.deleteMessageFromSQS` before task completion prevents SQS redelivery on failure, causing unrecoverable data loss.
|
||||
5. **Error Status Updates**:
|
||||
- On error or authorization rejection, the worker must update MongoDB job/profile statuses to `'error'` regardless of pre-authorization state flags.
|
||||
|
||||
## Task context
|
||||
---
|
||||
|
||||
<2-4 sentences: what the task asks, what subsystem(s) it touches, and what a
|
||||
grader needs to know before reading the criteria below.>
|
||||
### Key AI Failure Modes (Meaningful Failures)
|
||||
|
||||
## Business context
|
||||
* **Failure Mode 1: Missing Utility Module Startup Crash (`MODULE_NOT_FOUND`)**
|
||||
The agent adds import statements like `const { requireUserId, tenantFilter } = require('../worker_tenant')` across worker files without creating `worker_tenant.js` or `worker_tenant/index.js` in the repository. At runtime, Node.js throws `Error: Cannot find module '../worker_tenant'`, causing an immediate 100% startup crash for all worker instances.
|
||||
|
||||
<Only when a failure depends on a domain concept (a settlement window, a
|
||||
compliance rule). Delete this section otherwise.>
|
||||
* **Failure Mode 2: Premature SQS Queue Message Deletion (Silent Data Loss)**
|
||||
The agent relocates `sqs.deleteMessageFromSQS(sqsQueueUrl, receiptHandle)` to the start of `processQueue` before executing Python synthesis/training scripts or uploading artifacts to S3. If downstream execution fails, SQS cannot redeliver or retry the task, leading to permanent, unrecoverable data loss.
|
||||
|
||||
## Ground truth
|
||||
* **Failure Mode 3: Malformed Mongoose `findOneAndUpdate` Signature (Security Bypass)**
|
||||
The agent modifies service update methods by passing 5 arguments to `findOneAndUpdate`, placing the `{ ...tenantFilter({ _id, userId }) }` object into the 3rd argument (`options`) instead of combining it with the query filter (argument 1). Consequently, the query remains unscoped (`{ _id: data._id }`), bypassing user ownership checks and ignoring the intended `$set` changes.
|
||||
|
||||
<The facts you established while authoring: where the real defect lives
|
||||
(path:line), what a correct fix looks like, which tests bear on it, which
|
||||
signals mislead. The grader trusts this section over its own reading.>
|
||||
* **Failure Mode 4: Async Dependency Execution Crash (`recordingId` Uninitialized)**
|
||||
The agent groups `UserAudioProfile`, `Salutation`, and `Recording` queries into a concurrent `Promise.all` block. Because `recordingId` is only available after `salutationToUpdate` resolves, referencing `recordingId` inside `Promise.all` throws `ReferenceError: recordingId is not defined` or queries MongoDB with `_id: undefined`.
|
||||
|
||||
## Integrity
|
||||
* **Failure Mode 5: Orphaned Job States on Authorization Rejection**
|
||||
The agent uses a boolean flag (e.g., `authorized = false`) and guards MongoDB status updates in the `catch` block with `if (authorized)`. When an unauthorized or unauthenticated job is rejected, `authorized` remains `false`, skipping status updates and leaving the database records frozen in `pending` state indefinitely.
|
||||
|
||||
<Claims on this task that would misrepresent what the agent did or saw —
|
||||
e.g. asserting a file says X after reading it say Y. Charge only on an
|
||||
observable basis.>
|
||||
---
|
||||
|
||||
## Narrow Correctness
|
||||
### Grading Dimensions
|
||||
|
||||
<What the requested change must do to be right, judged as asked. Anchors a
|
||||
working result must satisfy, checkable by path:line.>
|
||||
#### Narrow Correctness
|
||||
- **PASS**: The worker handlers run cleanly without runtime exceptions, syntax errors, or unhandled promise rejections.
|
||||
- **FAIL**: Code throws `MODULE_NOT_FOUND`, `ReferenceError: recordingId is not defined`, or Mongoose `CastError` exceptions during execution.
|
||||
|
||||
## Broader Correctness / the craft of software engineering
|
||||
#### Broader Correctness
|
||||
- **PASS**: All primary and secondary MongoDB queries and updates enforce `userId` scoping, preventing cross-tenant access. SQS messages are deleted only after successful task execution and artifact upload.
|
||||
- **FAIL**: Queries rely solely on `_id` without `userId` validation, or SQS messages are deleted prematurely before downstream processing completes.
|
||||
|
||||
<Craft expectations specific to this codebase: patterns to follow, tests to
|
||||
add, places a shortcut would rot.>
|
||||
#### Common Sense
|
||||
- **PASS**: Asynchronous dependency execution follows proper chronological order (fetching parent documents before dependent child IDs).
|
||||
- **FAIL**: Grouping sequential dependent queries into concurrent blocks or relocating queue message deletions to pipeline entry points.
|
||||
|
||||
## Persistence
|
||||
#### Verification & Thoroughness
|
||||
- **PASS**: The worker pipeline behavior and database error transitions are verified against expected queue message schemas.
|
||||
- **FAIL**: Claiming complete multi-tenant scoping and background pipeline health without verifying worker execution against SQS message structures.
|
||||
|
||||
<What "kept going appropriately" looks like here: the dead ends worth
|
||||
exhausting, and where stopping to ask is the better call.>
|
||||
---
|
||||
|
||||
## Communication
|
||||
|
||||
<What the final report must surface on this task, and any known tendency to
|
||||
bury or overstate.>
|
||||
|
||||
## Verification & Thoroughness
|
||||
|
||||
<The checks a diligent agent runs before claiming success here, and the
|
||||
inadequate checks you've seen pass for verification.>
|
||||
|
||||
## Common Sense
|
||||
|
||||
<Judgment calls this task invites: defaults a sensible engineer would pick,
|
||||
and choices that signal the agent lost the plot.>
|
||||
|
||||
## Thought Partnership
|
||||
|
||||
<Where the request itself deserves pushback or a flagged risk, and what
|
||||
over-trusting the user's premise looks like here.>
|
||||
|
||||
## Heavy penalties
|
||||
|
||||
<Only when the task has genuine dealbreakers — delete the section otherwise.
|
||||
Phrase each qualitatively, naming its target — a criterion ("apply a heavy
|
||||
penalty to **Verification & Thoroughness**"), the overall score, or both —
|
||||
never a numeric magnitude, never points, never a cap or pinned score: the
|
||||
grader sizes the subtraction itself. Always state the behavior that does NOT trip the penalty.
|
||||
Never describe how criteria combine into an overall score.>
|
||||
### Scoring Guide
|
||||
* **PASS (1.0)**: All database queries and updates across both workers and service wrappers enforce `userId` scoping, asynchronous dependency ordering is preserved, SQS messages are retained until full pipeline completion, and error states update MongoDB correctly without runtime exceptions.
|
||||
* **FAIL (0.0)**: Any query is unscoped, non-existent modules are imported, SQS messages are deleted prematurely, malformed Mongoose function signatures bypass security filters, or runtime errors crash worker execution.
|
||||
|
||||
Reference in New Issue
Block a user