qvm: connect health interception

docs/ASSET_PIPELINE_TODO.md

@@ -889,3 +889,6 @@ Research progress: RE-740 [report](research/QVM_MODE10_SCORING.md); original ass
 
 
 Research progress: RE-741 [report](research/QVM_MODE10_COMPLETE.md); original assets unchanged.
+
+
+Research progress: RE-742 [report](research/QVM_HEALTH_MODE10.md); original assets unchanged.

docs/FINDINGS.md

@@ -1235,3 +1235,6 @@ RE-725–734 batch: [health, contact and Rust impulse findings](research/QVM_HEA
 
 
 **RE-741 — Complete mode10 interception source composition:** Complete206128 now composes entry gates, protected publication, descriptor dispatch, cleanup, obituary, counters and penalty requests across56 fixtures. [Evidence](research/QVM_MODE10_COMPLETE.md).
+
+
+**RE-742 — Actual lethal health suffix to complete mode10 interception:** The health lethal-decision suffix now executes actual206128; a blocked round still consumes the lethal intercept without ordinary subtraction. [Evidence](research/QVM_HEALTH_MODE10.md).

docs/PROVISIONAL_TODO.md

@@ -1224,3 +1224,6 @@ Research progress: RE-740 [report](research/QVM_MODE10_SCORING.md); full native/
 
 
 Research progress: RE-741 [report](research/QVM_MODE10_COMPLETE.md); full native/platform/game scope remains open.
+
+
+Research progress: RE-742 [report](research/QVM_HEALTH_MODE10.md); full native/platform/game scope remains open.

docs/RESEARCH_LEDGER.md

@@ -1351,3 +1351,10 @@ Interception increments victim/team counters, requests score+1 and same-team -2,
 ## 2026-10-21 — RE-741 Complete mode10 interception source composition
 
 Complete206128 now composes entry gates, protected publication, descriptor dispatch, cleanup, obituary, counters and penalty requests across56 fixtures. Focused `cargo xtask test-group mode10-complete` passes (3 Python tests); selected reports regenerate twice byte-identically, with storage, links and whitespace guards. New report: 547965 bytes. Full-suite validation is deferred to the tenth batch commit. No original assets, Rust runtime or shared VM evaluator changed.
+
+
+## 2026-10-21 — RE-742 Actual lethal health suffix to complete mode10 interception
+
+The health lethal-decision suffix now executes actual206128; a blocked round still consumes the lethal intercept without ordinary subtraction. Focused `cargo xtask test-group health-mode10` passes (3 Python tests); selected reports regenerate twice byte-identically, with storage, links and whitespace guards. New report: 456558 bytes. Full-suite validation is deferred to the tenth batch commit. Optional actual interception resolver added; prior default reports must remain byte-identical. Original assets and shared VM evaluator unchanged.
+
+Additional RE-742 focused health regression: `cargo xtask test-group health` passes71 Python tests and deterministic affected report regeneration; all prior tracked reports remain unchanged.

docs/REVERSE_ENGINEERING_TODO.md

@@ -1267,3 +1267,6 @@ Purpose: turn the supplied installation into an evidence-backed specification of
 
 
 - [x] RE-741 Complete mode10 interception source composition ([report](research/QVM_MODE10_COMPLETE.md)).
+
+
+- [x] RE-742 Actual lethal health suffix to complete mode10 interception ([report](research/QVM_HEALTH_MODE10.md)).

docs/research/QVM_HEALTH_MODE10.md

@@ -0,0 +1,7 @@
+# Actual lethal health suffix to complete mode10 interception
+
+RE-742. Reproduce with `python3 scripts/analyze_qvm_health_mode10.py`; validate with `cargo xtask test-group health-mode10`. See the [machine report](qvm-health-mode10.json).
+
+Composed212479..213006 with complete206128 and actual213009 over48 fixtures around equality, method6 exclusion, mode and round gates. Independent composed state/event/request oracles verify the child; ordinary subtraction remains checked on nonintercepting paths. The existing default captured handler stays compatible with prior reports. Earlier health gates/credit, child-native parity and actual score/lifecycle callback internals remain open.
+
+No original executable/full VM/native world collision or rendering runs. Supplied callback/native responses and fixture domains remain explicit; complete macOS/Linux Rust/Bevy/Solari gameplay remains open.

docs/research/README.md

@@ -825,3 +825,6 @@ The extension collector adds five focused tests for native tail calls, command-r
 
 
 [RE-741: Complete mode10 interception source composition](QVM_MODE10_COMPLETE.md).
+
+
+[RE-742: Actual lethal health suffix to complete mode10 interception](QVM_HEALTH_MODE10.md).

scripts/analyze_qvm_health_mode10.py

@@ -0,0 +1,18 @@
+#!/usr/bin/env python3
+"""Health lethal-decision suffix with actual206128 interception child."""
+import itertools
+from pathlib import Path
+from qvm_research_support import source
+from qvm_health_effects_support import seed as health_seed,run
+from qvm_lifecycle_support import seed as lifecycle_seed,A_CLIENT,TARGET,ATTACKER
+from analyze_qvm_health_impulse import impulse
+from analyze_qvm_mode10_complete import intercept
+from analyze_qvm import write_json
+
+def main():
+    q,raw,h,ins=source();rows=[]
+    for mode,health,damage,method,blocked in itertools.product([3,10],[99,100],[99,100],[6,12,34],[0,1]):
+        a=dict(mode=mode,health=health,damage=damage,method=method,blocked=blocked);m={**lifecycle_seed(raw,h,**a),**health_seed(raw,h,**a)};m.update({34992532:10000,34087532:1,34075396:3,A_CLIENT+3096:1,TARGET+168:65,ATTACKER+168:7})
+        _,p=run(ins,m,212479,213007,impulse_resolver=impulse,intercept_resolver=intercept);p['input']=a;rows.append(p)
+    write_json(Path('docs/research/qvm-health-mode10.json'),{'source':q['source'],'sha256':q['sha256'],'probes':rows,'limits':'Actual212479 lethal decision through health211239 suffix with actual206128 and213009. Earlier damage gates/credit not rerun; lifecycle allocations and unresolved child effects remain explicit stand-ins. Original runtime parity open.'})
+if __name__=='__main__':main()

scripts/qvm_health_effects_support.py

@@ -31,7 +31,7 @@ def state(m,target=TARGET):
     client=m[target+520]
     return {'entity':{str(o):m[target+o]&0xffffffff for o in ENTITY_FIELDS},'client':{str(o):m[client+o]&0xffffffff for o in CLIENT_FIELDS} if client else None,'attacker_credits':m.get(m.get(m.get(H+124,ATTACKER)+520,0)+252,0)&0xffffffff,'direction':[m[(m[H+128] or DIRECTION)+4*j]&0xffffffff for j in range(3)],'args':{str(o):m[H+o]&0xffffffff for o in [120,124,128,136,140,144,148]},'selected_damage':m[H+40]&0xffffffff,'knockback_word':m[H+32]&0xffffffff,'rng':m[21044]&0xffffffff}
 
-def handlers(effects,impulse_resolver=None):
+def handlers(effects,impulse_resolver=None,intercept_resolver=None):
     def root(i,m,a):v=word(math.sqrt(number(a[8])));return m,v,{'input_word':a[8]&0xffffffff,'result_word':v,'native_math_modeled':True}
     def rng(i,m,a):m,r=random_word(i,m);return m,r['result'],r
     def meter(i,m,a):
@@ -39,9 +39,13 @@ def handlers(effects,impulse_resolver=None):
     def impulse(i,m,a):
         if impulse_resolver:return impulse_resolver(i,m,a)
         return captured(213009,effects['impulses'])(i,m,a)
-    return {-107:root,33715:predicate,36717:rng,211197:meter,213009:impulse,206128:captured(206128,effects['intercepts']),152386:captured(152386,effects['protected_updates']),CALLBACK:captured(CALLBACK,effects['callbacks']),PAIN:captured(PAIN,effects['callbacks']),DEATH:captured(DEATH,effects['callbacks'])}
+    def intercepted(i,m,a):
+        if intercept_resolver:
+            m,value,r=intercept_resolver(i,m,a);effects['intercepts'].append({'arguments':a,'execution':r});return m,value,r
+        return captured(206128,effects['intercepts'])(i,m,a)
+    return {-107:root,33715:predicate,36717:rng,211197:meter,213009:impulse,206128:intercepted,152386:captured(152386,effects['protected_updates']),CALLBACK:captured(CALLBACK,effects['callbacks']),PAIN:captured(PAIN,effects['callbacks']),DEATH:captured(DEATH,effects['callbacks'])}
 
-def run(ins,m,start,boundary,impulse_resolver=None,target=TARGET):
+def run(ins,m,start,boundary,impulse_resolver=None,target=TARGET,intercept_resolver=None):
     before=state(m,target);effects={k:[] for k in ['callbacks','intercepts','protected_updates','impulses']}
-    m,e=resume(ins,start,213007,m,H,handlers=handlers(effects,impulse_resolver),boundaries=[] if boundary==213007 else [boundary],limit=50000)
+    m,e=resume(ins,start,213007,m,H,handlers=handlers(effects,impulse_resolver,intercept_resolver),boundaries=[] if boundary==213007 else [boundary],limit=50000)
     return m,{'initial':before,'final':state(m,target),'effects':effects,'execution':e}

scripts/test_qvm_health_mode10.py

@@ -0,0 +1,19 @@
+import json,unittest
+from pathlib import Path
+from test_qvm_mode10_complete import complete_expected
+class HealthMode10Tests(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):cls.ps=json.loads(Path('docs/research/qvm-health-mode10.json').read_text())['probes']
+    def test_lethal_decision_and_child_state_oracle(self):
+        for p in self.ps:
+            a=p['input'];selected=a['mode']==10 and a['health']-a['damage']<=0 and a['method']!=6;calls=p['effects']['intercepts'];self.assertEqual(len(calls),int(selected),a)
+            if selected:
+                child=calls[0]['execution'];r,targets=complete_expected({**child,'input':a},(0,0));self.assertEqual(child['final'],r);self.assertEqual([c['target'] for c in child['effects']['requests']],targets)
+    def test_interception_short_circuits_even_when_child_round_gate_blocks(self):
+        for p in self.ps:
+            a=p['input'];selected=bool(p['effects']['intercepts']);health=(a['health'] if a['blocked'] else 50) if selected else a['health']-a['damage'];self.assertEqual(p['final']['entity']['684'],health&0xffffffff,a)
+            if selected:self.assertEqual(p['effects']['impulses'],[]);self.assertEqual(p['effects']['callbacks'],[])
+    def test_exact_root_exit_and_domain(self):
+        self.assertEqual(len(self.ps),48)
+        for p in self.ps:self.assertEqual(p['execution']['stop'],213007)
+if __name__=='__main__':unittest.main()

xtask/checks.tsv

@@ -1188,3 +1188,6 @@ group	mode10-scoring	scripts/analyze_qvm_mode10_scoring.py
 report	scripts/analyze_qvm_mode10_complete.py	docs/research/qvm-mode10-complete.json
 edge	docs/research/qvm-mode10-complete.json	scripts/test_qvm_mode10_complete.py
 group	mode10-complete	scripts/analyze_qvm_mode10_complete.py
+report	scripts/analyze_qvm_health_mode10.py	docs/research/qvm-health-mode10.json
+edge	docs/research/qvm-health-mode10.json	scripts/test_qvm_health_mode10.py
+group	health-mode10	scripts/analyze_qvm_health_mode10.py