Bind the original Fremen/ally/turn grant before Karama allowance or prevention. A prevented grant leaves intrinsic free rates and prior payments intact. Permit late grants under the explicit provisional Recruits current-rate policy, with shared controls, minimal bot support and native public chronology evidence. Complete rules and browser/deployed acceptance remain open.
201 lines
12 KiB
Markdown
201 lines
12 KiB
Markdown
# Recruits development preview
|
||
|
||
14 September 2026. **Prototyped, not full card-variant acceptance.** The independent
|
||
Ecaz & Moritani Treachery Cards variant is included before genuine Basic or
|
||
Advanced faction setup. All three physical cards are present exactly once.
|
||
Recruits, bounded Reinforcements and Harass & Withdraw have connected controls,
|
||
AI and saved continuation; full interactions and publication remain gated.
|
||
|
||
## Native Fremen grant after play — 24 October 2026
|
||
|
||
The earlier later-grant guard is replaced by the symmetric current-rate
|
||
prototype: an original Fremen allied grant may be declared after Recruits.
|
||
Its native Karama window settles before permission is committed. Prevention
|
||
blocks only that grant for this Revival; it does not erase intrinsic free
|
||
allowance, Recruits or earlier returns/payments. A successful late grant uses
|
||
the same current3→6 calculation as a grant preceding play, with actual prior
|
||
free usage deducted, no refund, and ordinary/native total and typed limits.
|
||
The exact response scope and later-rate composition remain provisional under
|
||
14 October, not a new publisher combined ruling.
|
||
[Source, controls and bounded evidence](FREMEN_REVIVAL_GRANT.md).
|
||
|
||
## Original Skills/phase composition — 17 October 2026
|
||
|
||
Supported classic, selected E1/E2 and standalone/paired E3 all14/card profiles
|
||
retain original Tech3–6 Basic/Advanced and Advanced Strongholds2–6, without
|
||
World/Nexus/Discovery. One genuine five-Tanks programme returns3 for4, then2
|
||
for2 after Recruits, preserving prior payment and free/ordinary usage. The
|
||
larger allowance cannot create a third counter when only two actual Tanks
|
||
remain. Axlotl accrues once/pays at phase end; normal Banker keeps its actual
|
||
Bank-payer threshold and deferred Mentat grant. Native prices/caps/recipients
|
||
and provisional late-paid policy remain unchanged.
|
||
[Current471/471 programme and limits](LEADER_SKILLS_RUNTIME.md#independent-cards-with-original-phase-modules--17-october-2026).
|
||
|
||
## Provisional late-paid activation — 8 October 2026
|
||
|
||
The content-first pass now permits Recruits after recorded paid ordinary
|
||
returns, once the current transaction has finished. **Provisional policy:**
|
||
prior payments are retained; only actual recorded free returns consume the
|
||
doubled free rate, while all returns consume the total allowance. No refund,
|
||
reset or retroactive reclassification is performed. This is not a numerical
|
||
user ruling or a newly retrieved publisher answer.
|
||
|
||
Thus Atreides' actual two-free/one-paid return retains its two-spice payment
|
||
and three total returns; after Recruits it may take two additional free
|
||
forces under the seven total cap. Missing or inconsistent free-use ledgers,
|
||
pending declarations, unrelated later rate changes and same-turn replay remain guarded.
|
||
The owned before-play panel and public active panel name this interpretation.
|
||
|
||
Basic/Advanced before/after native smokes use a genuinely dealt card, an
|
||
actual three-force return and useful allied Easy activation, then two free
|
||
returns: total5/free4, unchanged paid spice and one card disposal. Live React
|
||
controls in Chromium produce the same available2/remaining4 quote and preserve
|
||
spice18/total3/free2 at activation. No broad suite, recovery or deployed proof.
|
||
|
||
|
||
## Classic Homeworld composition — 16 October 2026
|
||
|
||
The fresh `ecaz-treachery --homeworlds` entry now admits Recruits during
|
||
original classic2–6 Basic/Advanced Revival, with its original36-card inventory.
|
||
Reuse native current free rates and typed revival deposits. No rate snapshot,
|
||
new saved field, seven extra counters, automatic refund, generic module
|
||
activation or played-save conversion is added. Reinforcements has a separate
|
||
[native-source fee composition](REINFORCEMENTS_RUNTIME.md#classic-homeworld-reserve-fees--16-october-2026);
|
||
original Tech/Advanced Strongholds now follow the separate
|
||
[classic World phase contract](ECAZ_TREACHERY_RULES.md#classic-world-phase-module-composition--16-october-2026).
|
||
Selected E1/E2 Worlds use the [native setup contract](ECAZ_TREACHERY_RULES.md#native-selected-family-world-composition--17-october-2026)
|
||
and [original phase-module composition](ECAZ_TREACHERY_RULES.md#native-world-phase-modules--17-october-2026);
|
||
other rosters and overlays remain separate.
|
||
|
||
Publisher Ecaz & Moritani printed pp11/16 specifically gives low Fremen4,
|
||
doubled but capped at7. Under the standing choose-recommended instruction,
|
||
compose the existing native separate-request population rule: quote against
|
||
current population when each request is accepted, retaining that group's
|
||
accepted free/paid classification. Later separate requests observe the
|
||
resulting population and actual recorded free usage. This changing-threshold
|
||
timing is visibly provisional, not a new publisher worked example or answer
|
||
to the earlier late-rate question.
|
||
|
||
Thus low Fremen at2 native counters can accept7 free together. If it accepts1
|
||
first, reaching high3, the next native rate is3→6 and only5 free remain. With
|
||
Tleilaxu absent it cannot buy a seventh after those six. The total card cap
|
||
is still7, not an independent paid-revival permission. Original special-force
|
||
limits remain: Advanced one Fedaykin/Sardaukar per turn; Basic preserves
|
||
physical star identities without adding the Advanced one-per-turn rule.
|
||
Original destination/deployment eligibility remains native, not guaranteed
|
||
merely because Recruits increased a rate.
|
||
|
||
Atreides with actual3 prior returns (2 free/1 paid) can accept2 further free
|
||
after play, even if already ready while phase4 remains open. Its payment
|
||
stays2 spice, total5/free4. A native mixed Emperor return deposits normal on
|
||
Kaitain and Sardaukar on Salusa in Advanced; Basic retains its primary World.
|
||
The new panel names the provisional current-rate policy. An actual Advanced
|
||
Fedaykin return exposed an enabled repeat with stale selected elite1 against
|
||
remaining0: the button now disables that over-cap selection and explains the
|
||
remaining special limit, without altering authoritative revival rules.
|
||
|
||
Four actual390px GameTable journeys cover Basic bulk7, Advanced typed1+5,
|
||
already-ready paid Atreides and Advanced mixed Emperor revival. Native low→high
|
||
quotes, physical custody, no refund and exact once-only discard are observed.
|
||
Sixteen fresh programmes/four minimal policies accept104 card/Revival/readiness
|
||
actions through original phase5, including useful activation by an already-ready
|
||
Atreides. Real setup/shipments/zero-dial losses produce the dead counters;
|
||
only physical hands/unopened first-Blow cards are positioned, with the existing
|
||
Recruits moved to the selected holder explicitly. These are not unstaged
|
||
full-game card draws. Eight startup Vite dependency aborts and a corrected
|
||
QA-only middleware ordering issue precede successful surfaces; no overflow
|
||
or application exception. Owned hosts/tab/scaffolds removed; proof outside Git.
|
||
|
||
Final checks:10 new cases pass, related18 offline files167/167, types/lint and
|
||
app build pass. A test-only nullable-view narrowing is followed by a passing
|
||
10-case rerun; production behavior is unchanged. CLI help executes and guide
|
||
import resolves207 unique evidence paths/source anchor, with both coverages
|
||
Partial. Existing build register/large-chunk/route-classification warnings
|
||
remain; no HTTP/DB boundary changed or new recovery assurance ran.
|
||
|
||
## Rules contract
|
||
|
||
[The source audit](RECRUITS_RULES.md) records publisher Ecaz & Moritani pages 11
|
||
and 16, the base and faction revival rules, and designer Jack Reda's explanation
|
||
at 8:25–8:48. The card plays during Revival, doubles every current free rate,
|
||
raises the normal force limit to seven for this turn, and is discarded after use.
|
||
The publisher's four-to-eight example is capped at seven by the ordinary limit.
|
||
Fresh publisher/designer-linked PDF fetches on 14 September returned HTTP 403;
|
||
this follow-up uses the earlier recorded primary-source audit and independent
|
||
source review. No newly retrieved late-paid ruling is claimed.
|
||
|
||
The scope permits activation before returns or after consistently recorded
|
||
free/paid normal returns, once priority windows and transactions finish.
|
||
Unknown legacy usage, unrelated late-rate changes, same-turn replay after recovery
|
||
and other optional modules remain guarded, apart from clean classic Homeworlds
|
||
above or the separate clean classic/Ix/CHOAM
|
||
[Tech/Advanced Stronghold composition](ECAZ_TREACHERY_RULES.md#original-phase-module-composition--16-october-2026).
|
||
These are boundaries, not printed timing rules.
|
||
Already-ready players retain their Revival opportunity while phase four is open.
|
||
|
||
Atreides doubles two to four, Fremen three to six, and Guild one to two. Native
|
||
CHOAM and Tleilaxu retain their unlimited allowance with free rates zero and four.
|
||
Paid prices, prevention and separate elite caps remain in force. A Fremen grant
|
||
settled before play doubles the ally's replacement rate three to six.
|
||
|
||
The seven-force allowance remains an independent global card effect when a
|
||
CHOAM or Tleilaxu faction advantage is canceled. **This is an implementation
|
||
inference from independent effects, not a direct combined publisher FAQ.** A
|
||
Tleilaxu limit response is not opened merely for crossing three under Recruits;
|
||
other qualifying price/prevention responses retain their original purposes.
|
||
|
||
The earlier user question about one versus two further free Atreides returns
|
||
remains unanswered. The 8 October first version uses the visibly provisional
|
||
two-further-free policy above; it does not claim that question was adjudicated.
|
||
The source audit retains the alternatives and later changing-rate ambiguity.
|
||
|
||
## Connected path and persistence
|
||
|
||
`initializeEcazTreacheryGameForAudit` composes the complete three-card inventory
|
||
with the selected ordinary deck before shuffle/deal. The local `ecaz-treachery`
|
||
prototype-room profile accepts only a fresh ready lobby and uses the existing
|
||
version/state compare-and-swap. It preserves seats and other rooms.
|
||
|
||
The holder's existing inspector shows the complete original-language card guide.
|
||
The Revival panel submits the shared `recruitsPlayAction`; rivals receive no held
|
||
card offer. A successful action discards the physical card once and saves its
|
||
player, turn and identity. Every player then sees the same effective public rates
|
||
and force limits. No rival hand, private spice or revival-use ledger is added.
|
||
|
||
Authoritative shared revival quotes use the effect. Exact free usage is separate
|
||
from total normal returns only while Recruits is active; inactive/legacy games
|
||
retain their previous arithmetic. Active receipts and ledgers are validated on
|
||
action and projection. JSON and SQLite restoration retain the same allowance;
|
||
the next turn clears the effect. Rejected actions leave the supplied state intact.
|
||
|
||
All four AI profiles use the same legal owned offer when their faction or ally can
|
||
benefit, then ordinary quoted revival choices. The late Fremen grant is excluded
|
||
from bot candidates as well as server and UI controls. This is a legal first
|
||
strategy, not completed Recruits difficulty calibration.
|
||
|
||
## Verification
|
||
|
||
Focused engine, component, AI, genuine prototype-room and authenticated SQLite
|
||
recovery suites are maintained alongside the pure arithmetic and inventory tests.
|
||
Independent review passes a 37-case related union; four authenticated SQLite cases
|
||
cover restored free/paid returns, malformed saves, stale/duplicate actions and
|
||
competing writes. Eight fixed-seed six-seat Basic/Advanced samples across all four
|
||
AI profiles finish in 7,995 accepted actions with no rejected candidates. Recruits
|
||
is genuinely played three times. Physical cards, Traitors, safe force/elite pools,
|
||
212 periodic all-seat JSON restores and terminal private views pass. Undefined
|
||
object properties are normalized through JSON on both sides of the comparison;
|
||
defined values are retained. Earlier harness attempts remain in private artifacts.
|
||
|
||
A separate conserved browser Revival fixture verifies the full inspector, actual
|
||
play/discard, public Atreides four/Fremen six rates, four free returns, automatic
|
||
chronicle and active/resolved refresh. Exact state comparison preserves the seat,
|
||
cards, forces and unchanged spice. This staged browser position does not establish
|
||
naturally reached card use. Final broad checks, source fingerprints, saved-game
|
||
preservation and Git delivery are recorded with the checkpoint. These samples do
|
||
not certify the unfinished three-card variant or every combined configuration.
|
||
|
||
See [current status](CURRENT_STATUS.md), [decision index](RULE_DECISIONS.md) and
|
||
[the reference checklist](../game/reference.ts), topic `card-recruits`, for the
|
||
remaining boundaries. Final polishing, other two effects and valid optional-module
|
||
combinations remain required work.
|