chore: 181 - Add tests proving missing manifest attribution is rejected for

[Exohayvan]

Roadmap step

phase-1/step-181: Step 181 - Add tests proving missing manifest attribution is rejected for

Objective

Complete the selected roadmap step using only the programmatically selected step details.

Selected step

# Step 181 - Add tests proving missing manifest attribution is rejected for

## Source Checklist Item
- [ ] Add tests proving missing manifest attribution is rejected for expert/model jobs.

## Goal
Add focused tests that prove expert/model jobs cannot be accepted, executed, or recorded when their required manifest attribution is missing. The tests should protect creator node ID continuity, validation receipts, lineage, and contribution attribution by making absence of manifest data a hard failure instead of an implicit or best-effort path.

## Implementation Checklist
- [ ] Identify the job submission or execution boundary where expert/model manifest attribution is required.
- [ ] Add a negative test for an expert job with no manifest attribution.
- [ ] Add a negative test for a model job with no manifest attribution.
- [ ] Assert the job is rejected before contribution, lineage, or validation receipt records are created.
- [ ] Assert the rejection error is explicit enough to distinguish missing manifest attribution from generic validation failure.
- [ ] Keep fixtures local-first and deterministic, using only minimal valid job fields except the missing manifest attribution.

## Required Tests Or Verification
- [ ] Run the targeted attribution or job validation test suite locally.
- [ ] Verify both expert and model missing-manifest cases fail closed.
- [ ] Verify valid attributed expert/model fixtures still pass if covered by existing tests.
- [ ] Verify no creator node ID, manifest, lineage, validation receipt, or contribution attribution record is silently fabricated.

## Risks
- [ ] Tests may expose existing permissive behavior where jobs are accepted without attribution.
- [ ] Error assertions may become brittle if they depend on exact message text instead of stable error codes or types.
- [ ] Fixtures may accidentally omit unrelated required fields and test the wrong validation path.

## Done When
- [ ] Missing manifest attribution is covered for both expert and model jobs.
- [ ] Rejection happens before execution or contribution recording.
- [ ] The local test command passes with the new negative coverage included.

Changes

2c97a43 chore: complete phase-1 step-181
 tests/test_expert_manifest.py | 47 +++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 47 insertions(+)

Validation

• PYTHONDONTWRITEBYTECODE=1 python3 scripts/full_test.py --mode fast --base origin/main --keep-going

Safety notes

• Roadmap selection was performed by the Python loop, not by the AI agent.
• No force push used.
• No hard reset used.
• Local automation paths are intentionally omitted.