Dune/docs/EFFICIENCY_PILOT.md

Three-milestone efficiency pilot

Preserve the complete game scope and required checks. Use small relevant reads, coherent feature batches, selective subagents and compact verification artifacts. Git remains mandatory: review, verify, commit, push and confirm remote success at each finished implementation checkpoint. Never force-push or commit credentials. The removed recurring automation stays removed.

Milestone	Deliverable	Measurement
1: setup	Current status, decision index, reusable checks, saved-game snapshots/backups and HTTP seat restoration	Accepted checkpoint 6e26ba5. Aggregate counters 57,957,554 → 58,124,416: 166,862 tokens including planning, coordination and the user-requested save reset. Two bounded helpers. Fifteen added regressions; 4,177 offline tests passed. Three broad attempts: two stopped before tests on type/lint errors; final check passed in 130.147s. One cancellation defect found and fixed in review.
2: core readiness batch	Typed Advanced reserve-shipment commitments without Guild or optional modules	Accepted checkpoint 32791ed. Aggregate counters 58,124,416 → 58,695,053: 570,637 tokens. Two bounded helpers: regression implementation and independent rules/privacy review. Fifteen added tests; 4,192 offline and 40 HTTP tests plus build passed. One stable broad run; focused checks resolved setup mistakes and two review findings before it. Two browser rooms verified distinct AI and human paths. Pending Guild settlement remains explicit.
3: next dependency batch	Truthtrance total hand size and primary-role counts, controls, bots and recovery	Accepted checkpoint 7454456. Aggregate counter 58,695,053 → 59,144,653 at completed verification: 449,600 tokens, before final Git/push overhead. Two workhorse-model helpers for regression implementation and independent review. Reused count panel and SQLite fixtures; eight added tests and expanded recovery cases. Two copy issues fixed before the single stable broad run: 4,200 offline tests, build and 40 HTTP tests passed. One browser room covered combined roles, total size, invalid input and phone refresh.

The aggregate counter may update only between turns. Record a confirmed endpoint when available rather than estimating missing usage. It includes coordination and conversation overhead and is not a dollar-cost meter. Different capability sizes limit direct comparisons; do not promise a savings percentage from this setup. Do not count tests alone as accepted faction or mode completion. Scripts consolidate existing checks; broad checks still run once source/review is stable and repeat if subsequent changes or failures justify them. Browser fixtures remain feature-specific; the reusable HTTP restoration helper does not automate browser setup or visual QA.

Milestone 2 cost more than the tooling baseline and cannot establish a saving. Its larger rules/interaction scope, browser setup and context recovery contribute to that total. Milestone 3 was smaller and reused more existing structure; its lower count cannot isolate the effect of model choice or establish a savings rate. Avoid repeatedly loading old audits. Three different deliverables alone cannot support a reliable percentage estimate.

For initial planning, allow roughly 450,000–575,000 aggregate tokens for a similarly sized reviewed gameplay batch, based on milestones 2 and 3. This is a low-confidence planning estimate, not a dollar quote, token-budget setting or a reason to stop required work. Larger interaction sets need a revised estimate. The tooling baseline was about 167,000 and is not a gameplay cost target. Reassess after the next few comparable batches, including defects found after acceptance.

Next adjustment: use the owner plus one bounded helper for source research and independent review when the existing implementation needs verification. Reuse the same helper for its follow-up review, existing fixtures and applicable checks. Avoid adding runtime code merely to make an audit appear to be implementation. The full-plan interaction audit starts at aggregate counter 59,151,671, uses that one-helper pattern and changes tests/documentation. It therefore requires the full offline check, but can reuse the unchanged application's earlier build; browser/restart acceptance supplies direct recovery evidence. Checkpoint 4a0b1ee passed all 4,205 offline tests in one stable broad check. Its aggregate endpoint was 59,509,237 at completed verification: 357,566 tokens before final Git and new goal-steering overhead. It is an audit batch, not a new gameplay capability.

User-directed rapid prototyping, 13 September 2026

The first batch starts at aggregate counter 59,524,858 and connects genuine Ix setup to the existing faction/card gameplay and saved browser table. One bounded helper inventoried existing work and implemented focused tests; a second reviewed setup privacy and the local database writer. The owner integrates the prototype entry, reusable backed-up starter, checklist and browser journey. Six added tests include two fixed-seed complete games without a calibration study. A saved-version review finding and a test typing error were fixed before the stable broad check; final results and counter endpoint are recorded in the checkpoint commit.

Prioritize functional coverage of the entire remaining game before exhaustive per-feature refinement. Use the existing checklist to distinguish missing, prototyped, integrated, verified and polished functions, and choose connected batches by dependency across core and expansion systems. A prototype includes a working action, usable controls, legal AI participation and saved continuation. Track functions newly playable and issues discovered during integration, alongside rework; do not count placeholders as delivered functions. Earlier token counters remain historical measurements, not a required per-checkpoint task.

Keep focused invariant and smoke checks during prototyping and required checks at stable checkpoints. Use integrated games to find defects; defer exhaustive combination testing and final visual polish to refinement unless an active defect requires them sooner. Full AI implementation, strategic refinement and calibration wait until all non-AI features are complete; earlier AI work is limited to minimal legal participation and critical correctness fixes. Then pursue the three approximately 75% adjacent-pair targets in the AI plan. Preserve material rule questions, mode gates, selective subagents and automatic verified Git pushes. The earlier batch estimates remain historical planning evidence and should be reassessed for this broader prototyping workflow, without an assumed savings percentage.

On 20 September the user removed the prior usage cutoff and routine account-usage checks. Preserve Git and subagent requirements and leave the removed recurring automation absent. Earlier usage counters in this log are historical only.

Use current status, rule decisions and verification workflow as short entry points. Historical logs remain references, not mandatory full-context input on every turn.

The next rapid prototype batch begins at aggregate counter 59,982,405 and adds one missing action: CHOAM Nexus Secret Ally Collection trade. A workhorse helper acquires the complete Discovery inventory for the following batch; a reused independent reviewer checks this trade and its queued-Collection lock. The owner connects runtime, private controls, AI and SQLite recovery. The battle inspection alternative stays explicitly missing instead of holding the independent trade for an exhaustive battle-cleanup design. Final checks, counter endpoint and checkpoint scope are recorded in the commit.

The Discovery batch starts at aggregate counter 60,591,430. Three bounded helpers cover physical inventory, nested board/admission, and private controls/docs; reusing them adds deterministic playthroughs and independent review of the owner's encounter/reward integration. New playable functions are Great Maker's vote/reserve ride, six Discovery blows, private Collection inspection/reveal, and both immediate stash rewards with saved overflow discard. Token custody and board destinations prepare the following location/flight batch. Reused room-start tooling now selects the Discovery profile. Reviews found and repaired lost saved controls, a retargetable arrival and skipped overflow discard before the stable broad check. The implementation adds no calibration study or server restart. Final aggregate endpoint, source-bound checks and push belong in the checkpoint commit; whole-module completion remains unclaimed.

The following Discovery batch starts at aggregate counter 61,819,586 and connects three more functions: next-turn free entry, Cistern Collection income and later-turn Ornithopter movement. Three reused helpers build separate quotes and controls, then integrate Collection/flight and independently review entry/flight; the owner integrates entry, persistence tests and browser acceptance. Entry review repaired arrival custody and conflicting saved continuations; targeted browser play found an incorrect decision heading. Orgiz's unresolved stacked-blow interpretation remains disabled while the independent functions advance. One new browser room uses genuine setup followed by explicitly targeted scenarios; the 156 opening rooms are preserved. Required stable checks and the aggregate endpoint are recorded in the checkpoint commit, without counting these prototypes as complete Discovery rules compliance.

The next location-effects batch begins at aggregate counter 63,116,444. Two workhorse helpers implement Jacurutu and Shrine, then cross-review Shrine and review the owner's Testing Station. Reuse of the existing prototype starter, card-effect paths and SQLite fixtures connects three bounded functions. Review found a reserved-card Shrine deadlock, incomplete alternate-role permission and missing physical-card selection; these require fixes before broad verification. Station review strengthened restored source and physical-token binding. One focused rendered-control assertion was corrected to distinguish the native disabled attribute from the component's data-disabled attribute. A fresh browser scenario checks station controls and refresh; all 187 opening rooms remain the preservation baseline. The first broad run reached 4,345 offline tests with five failures: two exposed an unnecessary hand read before public Shrine entitlement, two old overlay fixtures queued nonexistent cards, and one assertion expected older rejection wording. The policy now checks public entitlement first; physical fixtures and wording are corrected without weakening custody validation. Focused regressions pass before the repeated stable broad run. Final aggregate endpoint and stable checks belong in the verified checkpoint commit. Pending rulings and complete-mode gates remain explicit rather than inflating accepted capability counts.