Resolve final panel launch conditions

[nedcut]

What changed

• reconcile the frozen publication protocol with the machine-enforced 4,096/3,072/8,192 native-minimum-reasoning lane
• add numeric protocol-to-lane regression assertions
• reserve every configured repair attempt as another full-price call and apply the committed 1.2x cost contingency before admitting a cell
• record the independent Fable 5 launch audit and the resulting $95 sub-$100 operating plan

Why

The final independent audit found no P0 launch blocker, but identified two pre-data conditions: the protocol document still described the retired 1,024-token policy, and the per-cell reservation covered only primary calls despite one bounded repair being configured. Both needed resolution before any full-panel result existed.

The strengthened reservation simulation uses accepted-smoke spend scaled to panel size. Expected total spend is $46.7742, and the largest expected measured-spend-plus-reservation commitment is $89.3659 before Mistral, so a $95 operator ceiling remains below the user's $100 limit.

Validation

• 459 Python tests passed
• Ruff format and lint passed
• benchmark contract validation passed
• web lint and production build passed
• panel dry-run expanded exactly 10 cells
• GM_BENCH_PRIVATE_SEEDS is unset after loading local env files
• final no-spend live preflight passed all 10 exact routes with the strengthened reservation policy

Summary by CodeRabbit

• Improvements

  • Publication runs now support native minimum reasoning with a 4,096-token cap and an 8,192-token fallback when cap pressure occurs.
  • Smoke validation requires successful completion at the standard cap before full-panel execution.
  • Cost reservations now include configured repair attempts and a 1.2× contingency for more reliable budget coverage.

• Documentation

  • Added final launch audit records, publication decision updates, spending estimates, operator limits, and monitoring requirements.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 45081a14-fcb6-4feb-a2a8-65f6b67fe713

📥 Commits

Reviewing files that changed from the base of the PR and between 0f12f21 and 8c8358c.

📒 Files selected for processing (6)

• config/publication_protocol.json
• docs/PUBLISH_READINESS.md
• docs/run_logs/sota-v2-final-launch-audit-2026-07-18.md
• scripts/run_publication_matrix.py
• tests/test_publication.py
• tests/test_publication_runner.py

---

📝 Walkthrough

Walkthrough

The publication protocol now uses native-minimum reasoning caps, while paid-cell reservations account for repair attempts and cost contingency. Tests validate protocol alignment and reservation calculations, and launch documentation records the resulting audit and operational conditions.

Changes

Publication readiness updates

Layer / File(s)	Summary
Revise publication protocol and consistency checks config/publication_protocol.json, tests/test_publication.py	The lane adopts 4,096-token output, 3,072-token pressure, 8,192-token fallback, native-minimum reasoning, updated smoke requirements, and matching configuration assertions.
Expand reservation accounting scripts/run_publication_matrix.py, tests/test_publication_runner.py	Paid-cell reservations include configured repair attempts and the pricing contingency multiplier, persist both values, and are validated for panel cells.
Record launch audit milestones docs/PUBLISH_READINESS.md, docs/run_logs/sota-v2-final-launch-audit-2026-07-18.md	Readiness records document the reconciled protocol, reservation contingency, spend estimates, operator ceiling, and final operational conditions.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Sequence Diagram(s)

sequenceDiagram
  participant PanelCell
  participant ReservationCalculator
  participant PricingSnapshot
  participant ReservationStore
  PanelCell->>ReservationCalculator: provide pricing and repair-attempt settings
  ReservationCalculator->>PricingSnapshot: read cost contingency multiplier
  PricingSnapshot-->>ReservationCalculator: return planning assumptions
  ReservationCalculator->>ReservationStore: persist amount and reservation metadata

Loading

Possibly related PRs

• nedcut/gm-bench#61 — Related publication cost, reservation, and validation work.
• nedcut/gm-bench#65 — Adds protocol repair-attempt metadata used by this reservation logic.
• nedcut/gm-bench#70 — Overlaps with reservation retry bookkeeping in the publication runner.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly matches the PR’s main goal of resolving final panel launch conditions.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch agent/final-panel-launch-gates

---

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.}

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.