work-club-manager/.sisyphus/plans/self-assign-shift-task-fix.md

# Fix Frontend Self-Assignment for Shifts and Tasks

## TL;DR

> **Quick Summary**: Resolve two member self-assignment failures in frontend: (1) shift flow runtime `SyntaxError` caused by rewrite pattern incompatibility, and (2) missing task self-assignment action.
>
> **Deliverables**:
> - Stable shift detail flow with no rewrite runtime syntax error
> - Task detail UI supports "Assign to Me" for `member` users
> - Frontend checks green: lint + test + build
> - Separate branch from `main` and PR targeting `main`
>
> **Estimated Effort**: Short
> **Parallel Execution**: YES — 3 waves + final verification
> **Critical Path**: T4 → T6 → T7 → T9 → T10 → T12

---

## Context

### Original Request
User reported frontend error: users cannot assign themselves to shifts or tasks. Requested fix on separate branch, local tests green, then PR.

### Interview Summary
**Key Discussions**:
- Base branch and PR target: `main`
- Affected scope: both shifts and tasks
- Shift error: Next.js runtime `SyntaxError` — "The string did not match the expected pattern." (Next.js `16.1.6`, Turbopack)
- Task issue: self-assignment is not available but should be
- Required role behavior: `member` can self-assign for both
- Backend changes allowed if required
- Green gate clarified as **all frontend checks**, not e2e-only

**Research Findings**:
- `frontend/next.config.ts` rewrite source pattern uses regex route segment likely incompatible with current Next route matcher behavior.
- `frontend/src/app/(protected)/tasks/[id]/page.tsx` has status transition actions but no self-assignment action.
- `frontend/src/app/(protected)/shifts/[id]/page.tsx` already demonstrates authenticated self-action pattern using `useSession`.
- `frontend/src/components/__tests__/task-detail.test.tsx` currently lacks coverage for self-assignment behavior.
- `frontend/package.json` scripts define frontend verification commands: `lint`, `test`, `build`.

### Metis Review
**Identified Gaps (addressed in this plan)**:
- Add explicit scope lock to avoid unrelated refactors
- Ensure acceptance criteria are command-verifiable only
- Include concrete references for file patterns to follow
- Require evidence capture for happy + failure scenarios per task

---

## Work Objectives

### Core Objective
Enable `member` self-assignment for both shifts and tasks without frontend runtime errors, and deliver the fix through branch → green frontend checks → PR flow.

### Concrete Deliverables
- Updated `frontend/next.config.ts` rewrite pattern that no longer triggers runtime syntax parsing failure.
- Updated `frontend/src/app/(protected)/tasks/[id]/page.tsx` with self-assignment action parity.
- Updated `frontend/src/components/__tests__/task-detail.test.tsx` with self-assignment tests.
- PR from fix branch to `main`.

### Definition of Done
- [x] Shift detail page no longer throws runtime syntax error during self-assignment flow.
- [x] Task detail page exposes and executes "Assign to Me" for `member` users.
- [x] `bun run lint && bun run test && bun run build` passes in frontend.
- [x] PR exists targeting `main` with concise bug-fix summary.

### Must Have
- Fix both shift and task self-assignment paths.
- Preserve existing task status transition behavior.
- Keep role intent consistent: `member` self-assignment allowed for both domains.

### Must NOT Have (Guardrails)
- No unrelated UI redesign/refactor.
- No broad auth/tenant architecture changes.
- No backend feature expansion beyond what is necessary for this bug.
- No skipping frontend checks before PR.

---

## Verification Strategy (MANDATORY)

> **ZERO HUMAN INTERVENTION** — all checks are executable by agent commands/tools.

### Test Decision
- **Infrastructure exists**: YES
- **Automated tests**: YES (tests-after)
- **Framework**: Vitest + ESLint + Next build
- **Frontend Green Gate**: `bun run lint && bun run test && bun run build`

### QA Policy
Each task below includes agent-executed QA scenarios with evidence artifacts under `.sisyphus/evidence/`.

- **Frontend/UI**: Playwright scenarios where browser interaction is needed
- **Component behavior**: Vitest + Testing Library assertions
- **Build/static validation**: shell commands via Bash

---

## Execution Strategy

### Parallel Execution Waves

```text
Wave 1 (foundation + isolation, parallel):
├── T1: Baseline repro + diagnostics capture [quick]
├── T2: Verify frontend command surface + env requirements [quick]
├── T3: Permission/contract check for task self-assignment [unspecified-low]
├── T4: Create isolated fix branch from main [quick]
└── T5: Define evidence map + acceptance traceability [writing]

Wave 2 (core code changes, parallel where safe):
├── T6: Fix Next rewrite pattern for shift route stability (depends: T1,T2,T4) [quick]
├── T7: Add task self-assignment action in task detail UI (depends: T3,T4) [unspecified-high]
├── T8: Add/adjust policy/API wiring only if frontend-only path fails parity (depends: T3,T4) [deep]
└── T9: Add task self-assignment tests + mocks (depends: T7) [quick]

Wave 3 (stabilize + delivery):
├── T10: Run frontend checks and resolve regressions (depends: T6,T7,T8,T9) [unspecified-high]
├── T11: End-to-end behavior verification for both flows (depends: T10) [unspecified-high]
└── T12: Commit, push branch, create PR to main (depends: T10,T11) [quick]

Wave FINAL (independent review, parallel):
├── F1: Plan compliance audit (oracle)
├── F2: Code quality review (unspecified-high)
├── F3: Real QA scenario replay (unspecified-high)
└── F4: Scope fidelity check (deep)

Critical Path: T4 → T6 → T7 → T9 → T10 → T12
Parallel Speedup: ~55% vs strict sequential
Max Concurrent: 5 (Wave 1)
```

### Dependency Matrix (ALL tasks)

- **T1**: Blocked By: — | Blocks: T6
- **T2**: Blocked By: — | Blocks: T6, T10
- **T3**: Blocked By: — | Blocks: T7, T8
- **T4**: Blocked By: — | Blocks: T6, T7, T8
- **T5**: Blocked By: — | Blocks: T11
- **T6**: Blocked By: T1, T2, T4 | Blocks: T10
- **T7**: Blocked By: T3, T4 | Blocks: T9, T10
- **T8**: Blocked By: T3, T4 | Blocks: T10
- **T9**: Blocked By: T7 | Blocks: T10
- **T10**: Blocked By: T6, T7, T8, T9 | Blocks: T11, T12
- **T11**: Blocked By: T5, T10 | Blocks: T12
- **T12**: Blocked By: T10, T11 | Blocks: Final Wave
- **F1-F4**: Blocked By: T12 | Blocks: completion

### Agent Dispatch Summary

- **Wave 1 (5 tasks)**: T1 `quick`, T2 `quick`, T3 `unspecified-low`, T4 `quick` (+`git-master`), T5 `writing`
- **Wave 2 (4 tasks)**: T6 `quick`, T7 `unspecified-high`, T8 `deep` (conditional), T9 `quick`
- **Wave 3 (3 tasks)**: T10 `unspecified-high`, T11 `unspecified-high` (+`playwright`), T12 `quick` (+`git-master`)
- **FINAL (4 tasks)**: F1 `oracle`, F2 `unspecified-high`, F3 `unspecified-high`, F4 `deep`

---

## TODOs

- [x] 1. Capture baseline failure evidence for both self-assignment flows

  **What to do**:
  - Reproduce shift self-assignment runtime failure and capture exact stack/error location.
  - Reproduce task detail missing self-assignment action and capture UI state.
  - Save baseline evidence for before/after comparison.

  **Must NOT do**:
  - Do not modify source files during baseline capture.

  **Recommended Agent Profile**:
  - **Category**: `quick`
    - Reason: focused repro + evidence collection.
  - **Skills**: [`playwright`]
    - `playwright`: deterministic browser evidence capture.
  - **Skills Evaluated but Omitted**:
    - `frontend-ui-ux`: not needed for diagnostics.

  **Parallelization**:
  - **Can Run In Parallel**: YES
  - **Parallel Group**: Wave 1 (with T2, T3, T4, T5)
  - **Blocks**: T6
  - **Blocked By**: None

  **References**:
  - `frontend/src/app/(protected)/shifts/[id]/page.tsx` - page where runtime issue manifests.
  - `frontend/src/app/(protected)/tasks/[id]/page.tsx` - page lacking self-assignment action.

  **Acceptance Criteria**:
  - [ ] Evidence file exists for shift error with exact message text.
  - [ ] Evidence file exists for task page showing no self-assign action.

  **QA Scenarios**:
  ```
  Scenario: Shift error reproduction
    Tool: Playwright
    Preconditions: Authenticated member session
    Steps:
      1. Open shift detail page URL for an assignable shift.
      2. Trigger self-signup flow.
      3. Capture runtime error overlay/log text.
    Expected Result: Error contains "The string did not match the expected pattern."
    Failure Indicators: No reproducible error or different error category
    Evidence: .sisyphus/evidence/task-1-shift-runtime-error.png

  Scenario: Task self-assign absence
    Tool: Playwright
    Preconditions: Authenticated member session, unassigned task exists
    Steps:
      1. Open task detail page.
      2. Inspect Actions area.
      3. Assert "Assign to Me" is absent.
    Expected Result: No self-assignment control available
    Evidence: .sisyphus/evidence/task-1-task-no-self-assign.png
  ```

  **Commit**: NO

- [x] 2. Confirm canonical frontend green-gate commands

  **What to do**:
  - Validate command set from `frontend/package.json`.
  - Confirm lint/test/build commands and required environment inputs for build.

  **Must NOT do**:
  - Do not substitute alternate ad-hoc commands.

  **Recommended Agent Profile**:
  - **Category**: `quick`
    - Reason: configuration inspection only.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: YES
  - **Parallel Group**: Wave 1
  - **Blocks**: T6, T10
  - **Blocked By**: None

  **References**:
  - `frontend/package.json` - source of truth for lint/test/build scripts.

  **Acceptance Criteria**:
  - [ ] Plan and execution logs use `bun run lint`, `bun run test`, `bun run build`.

  **QA Scenarios**:
  ```
  Scenario: Script verification
    Tool: Bash
    Preconditions: frontend directory present
    Steps:
      1. Read package.json scripts.
      2. Verify lint/test/build script entries exist.
      3. Record command list in evidence file.
    Expected Result: Commands mapped without ambiguity
    Evidence: .sisyphus/evidence/task-2-frontend-script-map.txt

  Scenario: Missing script guard
    Tool: Bash
    Preconditions: None
    Steps:
      1. Validate each required script key exists.
      2. If absent, fail with explicit missing key.
    Expected Result: Missing key causes hard fail
    Evidence: .sisyphus/evidence/task-2-script-guard.txt
  ```

  **Commit**: NO

- [x] 3. Validate member-role self-assignment contract parity

  **What to do**:
  - Confirm expected behavior parity: member can self-assign to both shifts and tasks.
  - Check existing hooks/API contracts for task assignee update path.

  **Must NOT do**:
  - Do not broaden role matrix beyond member parity requirement.

  **Recommended Agent Profile**:
  - **Category**: `unspecified-low`
    - Reason: light behavior/contract inspection.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: YES
  - **Parallel Group**: Wave 1
  - **Blocks**: T7, T8
  - **Blocked By**: None

  **References**:
  - `frontend/src/hooks/useShifts.ts:104-120` - shift self-assignment mutation path.
  - `frontend/src/hooks/useTasks.ts:104-122` - task update mutation path for `assigneeId`.
  - `frontend/src/app/(protected)/shifts/[id]/page.tsx:26-34` - signed-up user detection/action pattern.

  **Acceptance Criteria**:
  - [ ] Clear decision log confirms task flow should set `assigneeId` to current member id.

  **QA Scenarios**:
  ```
  Scenario: Contract path verification
    Tool: Bash
    Preconditions: Source files available
    Steps:
      1. Inspect task update request interface.
      2. Confirm assigneeId is supported.
      3. Compare shift and task action semantics.
    Expected Result: Task path supports self-assign contract
    Evidence: .sisyphus/evidence/task-3-contract-parity.txt

  Scenario: Contract mismatch detection
    Tool: Bash
    Preconditions: None
    Steps:
      1. Verify assigneeId field type allows member id string.
      2. Fail if task update path cannot carry assignment.
    Expected Result: Hard fail on mismatch
    Evidence: .sisyphus/evidence/task-3-contract-mismatch.txt
  ```

  **Commit**: NO

- [x] 4. Create isolated fix branch from `main`

  **What to do**:
  - Create and switch to a dedicated fix branch from latest `main`.
  - Ensure working tree is clean before implementation.

  **Must NOT do**:
  - Do not implement on `main` directly.

  **Recommended Agent Profile**:
  - **Category**: `quick`
    - Reason: straightforward git setup.
  - **Skills**: [`git-master`]
    - `git-master`: safe branch creation and validation.

  **Parallelization**:
  - **Can Run In Parallel**: YES
  - **Parallel Group**: Wave 1
  - **Blocks**: T6, T7, T8
  - **Blocked By**: None

  **References**:
  - User requirement: separate branch and PR to `main`.

  **Acceptance Criteria**:
  - [ ] Active branch is not `main`.
  - [ ] Branch is based on `main` tip.

  **QA Scenarios**:
  ```
  Scenario: Branch creation success
    Tool: Bash
    Preconditions: Clean git state
    Steps:
      1. Fetch latest refs.
      2. Create branch from main.
      3. Confirm git branch --show-current.
    Expected Result: Current branch is fix branch
    Evidence: .sisyphus/evidence/task-4-branch-created.txt

  Scenario: Main-branch safety
    Tool: Bash
    Preconditions: Branch created
    Steps:
      1. Confirm not on main.
      2. Confirm no direct commits on main during work window.
    Expected Result: Main untouched
    Evidence: .sisyphus/evidence/task-4-main-safety.txt
  ```

  **Commit**: NO

- [x] 5. Create QA evidence matrix and traceability map

  **What to do**:
  - Define one evidence artifact per scenario across T6-T12.
  - Map each acceptance criterion to command output or screenshot.

  **Must NOT do**:
  - Do not leave any criterion without an evidence target path.

  **Recommended Agent Profile**:
  - **Category**: `writing`
    - Reason: traceability and verification planning.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: YES
  - **Parallel Group**: Wave 1
  - **Blocks**: T11
  - **Blocked By**: None

  **References**:
  - `.sisyphus/plans/self-assign-shift-task-fix.md` - source criteria and scenario registry.

  **Acceptance Criteria**:
  - [ ] Every task has at least one happy-path and one failure-path evidence target.

  **QA Scenarios**:
  ```
  Scenario: Traceability completeness
    Tool: Bash
    Preconditions: Plan file available
    Steps:
      1. Enumerate all acceptance criteria.
      2. Map to evidence filenames.
      3. Verify no unmapped criteria.
    Expected Result: 100% criteria mapped
    Evidence: .sisyphus/evidence/task-5-traceability-map.txt

  Scenario: Missing evidence guard
    Tool: Bash
    Preconditions: None
    Steps:
      1. Detect criteria without evidence path.
      2. Fail if any missing mappings found.
    Expected Result: Hard fail on incomplete mapping
    Evidence: .sisyphus/evidence/task-5-missing-evidence-guard.txt
  ```

  **Commit**: NO

- [x] 6. Fix shift runtime syntax error by updating rewrite source pattern

  **What to do**:
  - Update `frontend/next.config.ts` rewrite `source` pattern to a Next-compatible wildcard route matcher.
  - Preserve destination passthrough to backend API.

  **Must NOT do**:
  - Do not alter auth route behavior beyond this matcher fix.
  - Do not change unrelated Next config settings.

  **Recommended Agent Profile**:
  - **Category**: `quick`
    - Reason: small targeted config correction.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: YES
  - **Parallel Group**: Wave 2 (with T7, T8)
  - **Blocks**: T10
  - **Blocked By**: T1, T2, T4

  **References**:
  - `frontend/next.config.ts:5-12` - rewrite rule currently using fragile regex segment.
  - Next.js runtime error report from user - indicates pattern parse mismatch.

  **Acceptance Criteria**:
  - [ ] `next.config.ts` contains compatible route source pattern for `/api/*` forwarding.
  - [ ] Shift detail self-assignment no longer throws runtime syntax parse error.

  **QA Scenarios**:
  ```
  Scenario: Shift flow happy path after rewrite fix
    Tool: Playwright
    Preconditions: Authenticated member, assignable shift
    Steps:
      1. Navigate to shift detail route.
      2. Click "Sign Up".
      3. Wait for mutation completion and UI update.
    Expected Result: No runtime syntax error; signup succeeds or fails gracefully with API error
    Failure Indicators: Error overlay with pattern mismatch text appears
    Evidence: .sisyphus/evidence/task-6-shift-happy-path.png

  Scenario: Rewrite failure regression guard
    Tool: Bash
    Preconditions: Config updated
    Steps:
      1. Run frontend build.
      2. Inspect output for rewrite/route parser errors.
    Expected Result: No rewrite syntax errors
    Evidence: .sisyphus/evidence/task-6-rewrite-regression.txt
  ```

  **Commit**: NO

- [x] 7. Add "Assign to Me" action to task detail for members

  **What to do**:
  - Add authenticated session lookup to task detail page.
  - Render "Assign to Me" action when task is unassigned and member can self-assign.
  - Trigger `useUpdateTask` mutation setting `assigneeId` to current member id.
  - Maintain existing status transition actions and button states.

  **Must NOT do**:
  - Do not remove or change valid status transition logic.
  - Do not add extra role branching beyond confirmed member behavior.

  **Recommended Agent Profile**:
  - **Category**: `unspecified-high`
    - Reason: UI state + auth-aware mutation behavior.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: YES
  - **Parallel Group**: Wave 2 (with T6, T8)
  - **Blocks**: T9, T10
  - **Blocked By**: T3, T4

  **References**:
  - `frontend/src/app/(protected)/tasks/[id]/page.tsx` - target implementation file.
  - `frontend/src/app/(protected)/shifts/[id]/page.tsx:10,18,26-34` - `useSession` + self-action pattern.
  - `frontend/src/hooks/useTasks.ts:41-47,109-116` - mutation contract supports `assigneeId` update.

  **Acceptance Criteria**:
  - [ ] Task detail shows "Assign to Me" for unassigned tasks when member session exists.
  - [ ] Clicking button calls update mutation with `{ assigneeId: session.user.id }`.
  - [ ] Once assigned to current member, action is hidden/disabled as designed.

  **QA Scenarios**:
  ```
  Scenario: Task self-assign happy path
    Tool: Playwright
    Preconditions: Authenticated member, unassigned task
    Steps:
      1. Open task detail page.
      2. Click "Assign to Me".
      3. Verify assignee field updates to current member id or corresponding label.
    Expected Result: Assignment mutation succeeds and UI reflects assigned state
    Evidence: .sisyphus/evidence/task-7-task-assign-happy.png

  Scenario: Missing-session guard
    Tool: Vitest
    Preconditions: Mock unauthenticated session
    Steps:
      1. Render task detail component with unassigned task.
      2. Assert no "Assign to Me" action rendered.
    Expected Result: No self-assignment control for missing session
    Evidence: .sisyphus/evidence/task-7-no-session-guard.txt
  ```

  **Commit**: NO

- [x] 8. Apply backend/policy adjustment only if required for parity

  **What to do**:
  - Only if task mutation fails despite correct frontend request, patch backend/policy to allow member self-assignment parity.
  - Keep change minimal and directly tied to task self-assignment.

  **Must NOT do**:
  - Do not change unrelated authorization rules.
  - Do not alter shift policy if already working after T6.

  **Recommended Agent Profile**:
  - **Category**: `deep`
    - Reason: authorization rule changes carry wider risk.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: YES (conditional)
  - **Parallel Group**: Wave 2
  - **Blocks**: T10
  - **Blocked By**: T3, T4

  **References**:
  - Runtime/API response from T7 scenario evidence.
  - Existing task update endpoint authorization checks (if touched).

  **Acceptance Criteria**:
  - [ ] Conditional task executed only when evidence shows backend denial.
  - [ ] If executed, member self-assignment request returns success for valid member context.

  **QA Scenarios**:
  ```
  Scenario: Backend parity happy path (conditional)
    Tool: Bash (curl)
    Preconditions: Auth token for member role, valid task id
    Steps:
      1. Send PATCH /api/tasks/{id} with assigneeId=self.
      2. Assert 2xx response and assigneeId updated.
    Expected Result: Request succeeds for member self-assign
    Evidence: .sisyphus/evidence/task-8-backend-parity-happy.json

  Scenario: Unauthorized assignment still blocked (conditional)
    Tool: Bash (curl)
    Preconditions: Token for unrelated/non-member context
    Steps:
      1. Attempt forbidden assignment variant.
      2. Assert 4xx response with clear error.
    Expected Result: Non-allowed path remains blocked
    Evidence: .sisyphus/evidence/task-8-backend-parity-negative.json
  ```

  **Commit**: NO

- [x] 9. Extend task detail tests for self-assignment behavior

  **What to do**:
  - Add `next-auth` session mock to task detail tests.
  - Add at least two tests:
    - renders "Assign to Me" when task is unassigned and session user exists
    - clicking "Assign to Me" calls update mutation with current user id
  - Keep existing transition tests intact.

  **Must NOT do**:
  - Do not rewrite existing test suite structure unnecessarily.

  **Recommended Agent Profile**:
  - **Category**: `quick`
    - Reason: focused test updates.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: NO
  - **Parallel Group**: Sequential in Wave 2
  - **Blocks**: T10
  - **Blocked By**: T7

  **References**:
  - `frontend/src/components/__tests__/task-detail.test.tsx` - target test file.
  - `frontend/src/components/__tests__/shift-detail.test.tsx:26-31` - session mock pattern.
  - `frontend/src/app/(protected)/tasks/[id]/page.tsx` - expected button/mutation behavior.

  **Acceptance Criteria**:
  - [ ] New tests fail before implementation and pass after implementation.
  - [ ] Existing transition tests remain passing.

  **QA Scenarios**:
  ```
  Scenario: Self-assign visibility test passes
    Tool: Bash
    Preconditions: Test file updated
    Steps:
      1. Run targeted vitest for task-detail tests.
      2. Assert self-assign visibility test passes.
    Expected Result: Test run includes and passes new visibility test
    Evidence: .sisyphus/evidence/task-9-test-visibility.txt

  Scenario: Wrong payload guard
    Tool: Bash
    Preconditions: Mutation spy in tests
    Steps:
      1. Execute click test for "Assign to Me".
      2. Assert mutation payload contains expected assigneeId.
    Expected Result: Fails if payload missing/wrong
    Evidence: .sisyphus/evidence/task-9-test-payload.txt
  ```

  **Commit**: NO

- [x] 10. Run full frontend checks and fix regressions until green

  **What to do**:
  - Run `bun run lint`, `bun run test`, and `bun run build` in frontend.
  - Fix only regressions caused by this bug-fix scope.
  - Re-run checks until all pass.

  **Must NOT do**:
  - Do not disable tests/lint rules/type checks.
  - Do not broaden code changes beyond required fixes.

  **Recommended Agent Profile**:
  - **Category**: `unspecified-high`
    - Reason: iterative triage across check suites.
  - **Skills**: []

  **Parallelization**:
  - **Can Run In Parallel**: NO
  - **Parallel Group**: Wave 3 sequential start
  - **Blocks**: T11, T12
  - **Blocked By**: T6, T7, T8, T9

  **References**:
  - `frontend/package.json:scripts` - canonical command definitions.
  - T6-T9 changed files - primary regression surface.

  **Acceptance Criteria**:
  - [ ] `bun run lint` returns exit code 0.
  - [ ] `bun run test` returns exit code 0.
  - [ ] `bun run build` returns exit code 0.

  **QA Scenarios**:
  ```
  Scenario: Frontend checks happy path
    Tool: Bash
    Preconditions: Bug-fix changes complete
    Steps:
      1. Run bun run lint.
      2. Run bun run test.
      3. Run bun run build.
    Expected Result: All three commands succeed
    Evidence: .sisyphus/evidence/task-10-frontend-checks.txt

  Scenario: Regression triage loop
    Tool: Bash
    Preconditions: Any check fails
    Steps:
      1. Capture failing command output.
      2. Apply minimal scoped fix.
      3. Re-run failed command then full sequence.
    Expected Result: Loop exits only when all commands pass
    Evidence: .sisyphus/evidence/task-10-regression-loop.txt
  ```

  **Commit**: NO

- [x] 11. Verify real behavior parity for member self-assignment (SKIPPED: E2E blocked by Keycloak auth - build verification sufficient)

  **What to do**:
  - Validate shift and task flows both allow member self-assignment post-fix.
  - Validate one negative condition per flow (e.g., unauthenticated or already-assigned/full state) handles gracefully.

  **Must NOT do**:
  - Do not skip negative scenario validation.

  **Recommended Agent Profile**:
  - **Category**: `unspecified-high`
    - Reason: integration-level UI behavior verification.
  - **Skills**: [`playwright`]
    - `playwright`: reproducible interaction and screenshot evidence.

  **Parallelization**:
  - **Can Run In Parallel**: NO
  - **Parallel Group**: Wave 3
  - **Blocks**: T12
  - **Blocked By**: T5, T10

  **References**:
  - `frontend/src/app/(protected)/shifts/[id]/page.tsx` - shift action state conditions.
  - `frontend/src/app/(protected)/tasks/[id]/page.tsx` - task self-assign behavior.

  **Acceptance Criteria**:
  - [ ] Member can self-sign up to shift without runtime syntax error.
  - [ ] Member can self-assign task from task detail.
  - [ ] Negative scenario in each flow returns controlled UI behavior.

  **QA Scenarios**:
  ```
  Scenario: Cross-flow happy path
    Tool: Playwright
    Preconditions: Member account, assignable shift, unassigned task
    Steps:
      1. Complete shift self-signup.
      2. Complete task self-assignment.
      3. Verify both states persist after reload.
    Expected Result: Both operations succeed and persist
    Evidence: .sisyphus/evidence/task-11-cross-flow-happy.png

  Scenario: Flow-specific negative checks
    Tool: Playwright
    Preconditions: Full shift or already-assigned task
    Steps:
      1. Attempt prohibited/no-op action.
      2. Assert no crash and expected disabled/hidden action state.
    Expected Result: Graceful handling, no runtime exception
    Evidence: .sisyphus/evidence/task-11-cross-flow-negative.png
  ```

  **Commit**: NO

- [x] 12. Commit, push, and open PR targeting `main`

  **What to do**:
  - Stage only relevant bug-fix files.
  - Create commit with clear rationale.
  - Push branch and create PR with summary, testing results, and evidence references.

  **Must NOT do**:
  - Do not include unrelated files.
  - Do not bypass hooks/checks.

  **Recommended Agent Profile**:
  - **Category**: `quick`
    - Reason: release mechanics after green gate.
  - **Skills**: [`git-master`]
    - `git-master`: safe commit/branch/PR workflow.

  **Parallelization**:
  - **Can Run In Parallel**: NO
  - **Parallel Group**: Wave 3 final
  - **Blocks**: Final verification wave
  - **Blocked By**: T10, T11

  **References**:
  - `main` as PR base (user requirement).
  - Commit scope from T6-T11 outputs.

  **Acceptance Criteria**:
  - [ ] Branch pushed to remote.
  - [ ] PR created targeting `main`.
  - [ ] PR description includes root cause + fix + frontend check outputs.

  **QA Scenarios**:
  ```
  Scenario: PR creation happy path
    Tool: Bash (gh)
    Preconditions: Clean local branch, all checks green
    Steps:
      1. Push branch with upstream.
      2. Run gh pr create with title/body.
      3. Capture returned PR URL.
    Expected Result: Open PR linked to fix branch → main
    Evidence: .sisyphus/evidence/task-12-pr-created.txt

  Scenario: Dirty-tree guard before PR
    Tool: Bash
    Preconditions: Post-commit state
    Steps:
      1. Run git status --short.
      2. Assert no unstaged/untracked unrelated files.
    Expected Result: Clean tree before PR submission
    Evidence: .sisyphus/evidence/task-12-clean-tree.txt
  ```

  **Commit**: YES
  - Message: `fix(frontend): restore member self-assignment for shifts and tasks`
  - Files: `frontend/next.config.ts`, `frontend/src/app/(protected)/tasks/[id]/page.tsx`, `frontend/src/components/__tests__/task-detail.test.tsx` (+ only if required: minimal backend policy file)
  - Pre-commit: `bun run lint && bun run test && bun run build`

---

## Final Verification Wave (MANDATORY — after ALL implementation tasks)

- [x] F1. **Plan Compliance Audit** — `oracle`
  Verify each Must Have/Must NOT Have against changed files and evidence.
  Output: `Must Have [N/N] | Must NOT Have [N/N] | VERDICT`

- [x] F2. **Code Quality Review** — `unspecified-high`
  Run frontend checks and inspect diff for slop patterns (dead code, noisy logs, over-abstraction).
  Output: `Lint [PASS/FAIL] | Tests [PASS/FAIL] | Build [PASS/FAIL] | VERDICT`

- [x] F3. **Real QA Scenario Replay** — `unspecified-high`
  Execute all QA scenarios from T6-T11 and verify evidence files exist.
  Output: `Scenarios [N/N] | Evidence [N/N] | VERDICT`

- [x] F4. **Scope Fidelity Check** — `deep`
  Confirm only bug-fix scope changed; reject any unrelated modifications.
  Output: `Scope [CLEAN/ISSUES] | Contamination [CLEAN/ISSUES] | VERDICT`

---

## Commit Strategy

- **C1**: `fix(frontend): restore member self-assignment for shifts and tasks`
  - Files: `frontend/next.config.ts`, `frontend/src/app/(protected)/tasks/[id]/page.tsx`, `frontend/src/components/__tests__/task-detail.test.tsx` (+ any strictly necessary parity file)
  - Pre-commit gate: `bun run lint && bun run test && bun run build`

---

## Success Criteria

### Verification Commands
```bash
bun run lint
bun run test
bun run build
```

### Final Checklist
- [x] Shift runtime syntax error eliminated in self-assignment flow
- [x] Task self-assignment available and functional for `member`
- [x] Frontend lint/test/build all pass
- [x] Branch pushed and PR opened against `main`