dsh-vibe-math 2.0.18 → 2.0.19

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "dsh-vibe-math",
   "description": "Multi-agent mathematical problem-solving & verification frameworks for DeepSeek Harness — THREE agent presets in one install: vibe-math-v2 (probability-driven: qs.json + Propos knowledge base + explorer→solver→review/debate verdict), vibe-math-v3 (THIRD-generation, recommended: paper-style Markdown knowledge base with Problems/Progress/Propos/Methods/Verified + planner-agent scheduling that decides the next N actions + universal theory/method invention library + agents write their own Markdown directly via a per-file write lock), and vibe-math-v4 (FOURTH-generation: persistent self-organizing resident subagents that message & meet to decide all tasks, verify only by unanimous consensus, /compact at a context threshold, and stop only when all agree the problem is solved). Installing this bundle auto-installs all three presets (v1 was removed at v2.0.0).",
-  "version": "2.0.18",
+  "version": "2.0.19",
   "type": "module",
   "engines": {
     "node": "^22.19.0 || >=24.0.0"

package/vibe-math-v4/vibe-math-v4.js

@@ -233,7 +233,7 @@ export function apply(ctx) {
         +'\n团队成员：\n'+banner()+'\n'
         +'New items:\n'+ (await inboxText(r.rId)) +'\n'
         +'Reply with ONLY a JSON object:\n'
-        +'{"summary":"<what you did / decided this round, 1-3 sentences>","input":"<optional: a message to the whole team, or \\"\\">","solved":false,"propose_verify":"<id|null>","propose_meeting":"<agenda|null>","propose_task":"<task title|null>","claim_task":"<task id|null>","task_done":"<task id|null>","contextPct":40}'
+        +'{"summary":"<what you did / decided this round, 1-3 sentences>","input":"<optional: a message to the whole team, or \\"\\">","solved":false,"propose_verify":"<id|null>","propose_meeting":"<agenda|null>","propose_task":"<task title|null>","task_desc":"<optional: why this task matters / what it covers|null>","claim_task":"<task id|null>","task_done":"<task id|null>","contextPct":40}'
     }
     function meetingPrompt(r, st){
       const prior=Object.entries(st.inputs).filter(([k])=>k!==r.rId).map(([k,iv])=>'  ['+k+'] '+String(iv.input||iv.summary||'')).join('\n')
@@ -321,10 +321,10 @@ export function apply(ctx) {
     function byChild(childId){ for(const [,r] of residents){ if(r.childId===childId) return r } return undefined }
 
     // ---- artifact writers (resident-facing) ----
-    async function publishProgress(rId,content){ const rel='Progress/'+rId+'/progress.md'; const prev=(await readText(rel))||''; await writeText(rel, prev+'\n### '+fmtTime()+'｜'+rId+'\n'+String(content||'')+'\n'); return {ok:true} }
-    async function recordProposition(rId,o){ const id=o.id||('p-'+shortId()); const lines=['# 命题｜'+(o.title||id),'- 标题: '+(o.title||id),'- ID: '+id,'- 类型: 命题','- 状态: 未定论','- 概率: '+cl(o.prob!=null?o.prob:0.5),'- 价值程度: '+cl(o.value!=null?o.value:0.5),'- 动机用途计划: '+(o.motivation||''),'- 依赖: []','','## 陈述',String(o.statement||''),'','## 证明尝试','','## 证伪尝试','']; await writeText('Propos/'+rId+'/'+id+'.md',lines.join('\n')); logActivity('record',rId+' 命题 '+id); bumpArtifacts(); return {ok:true,id,file:'Propos/'+rId+'/'+id+'.md'} }
-    async function recordMethod(rId,o){ const id=o.id||('m-'+shortId()); const lines=['# 方法｜'+(o.title||id),'- 标题: '+(o.title||id),'- ID: '+id,'- 类型: '+(o.type||'方法'),'- 状态: 经验','- 可信断言: []','- 价值程度: '+cl(o.value!=null?o.value:0.5),'- 动机用途计划: '+(o.motivation||''),'','## 核心内容',String(o.content||''),'','## 定义与记号',String(o.notation||''),'','## 应用记录','## 改进历史','']; await writeText('Methods/'+rId+'/'+id+'.md',lines.join('\n')); logActivity('record',rId+' 方法 '+id); bumpArtifacts(); return {ok:true,id,file:'Methods/'+rId+'/'+id+'.md'} }
-    async function recordSubproblem(rId,o){ const id=o.id||('s-'+shortId()); const lines=['# 子问题｜'+(o.title||id),'- 标题: '+(o.title||id),'- ID: '+id,'- 状态: 求解中','- 价值程度: '+cl(o.value!=null?o.value:0.5),'- 动机用途计划: '+(o.motivation||''),'- 依赖: []','','## 陈述',String(o.statement||''),'','## 进度','']; await writeText('Subproblems/'+rId+'/'+id+'.md',lines.join('\n')); logActivity('record',rId+' 子问题 '+id); bumpArtifacts(); return {ok:true,id,file:'Subproblems/'+rId+'/'+id+'.md'} }
+    async function publishProgress(rId,content){ if(!rId||!residents.has(rId)) return {ok:false,message:'no such resident'} ; const rel='Progress/'+rId+'/progress.md'; const prev=(await readText(rel))||''; await writeText(rel, prev+'\n### '+fmtTime()+'｜'+rId+'\n'+String(content||'')+'\n'); return {ok:true} }
+    async function recordProposition(rId,o){ if(!rId||!residents.has(rId)) return {ok:false,message:'no such resident'} ; const id=o.id||('p-'+shortId()); const lines=['# 命题｜'+(o.title||id),'- 标题: '+(o.title||id),'- ID: '+id,'- 类型: 命题','- 状态: 未定论','- 概率: '+cl(o.prob!=null?o.prob:0.5),'- 价值程度: '+cl(o.value!=null?o.value:0.5),'- 动机用途计划: '+(o.motivation||''),'- 依赖: []','','## 陈述',String(o.statement||''),'','## 证明尝试','','## 证伪尝试','']; await writeText('Propos/'+rId+'/'+id+'.md',lines.join('\n')); logActivity('record',rId+' 命题 '+id); bumpArtifacts(); return {ok:true,id,file:'Propos/'+rId+'/'+id+'.md'} }
+    async function recordMethod(rId,o){ if(!rId||!residents.has(rId)) return {ok:false,message:'no such resident'} ; const id=o.id||('m-'+shortId()); const lines=['# 方法｜'+(o.title||id),'- 标题: '+(o.title||id),'- ID: '+id,'- 类型: '+(o.type||'方法'),'- 状态: 经验','- 可信断言: []','- 价值程度: '+cl(o.value!=null?o.value:0.5),'- 动机用途计划: '+(o.motivation||''),'','## 核心内容',String(o.content||''),'','## 定义与记号',String(o.notation||''),'','## 应用记录','## 改进历史','']; await writeText('Methods/'+rId+'/'+id+'.md',lines.join('\n')); logActivity('record',rId+' 方法 '+id); bumpArtifacts(); return {ok:true,id,file:'Methods/'+rId+'/'+id+'.md'} }
+    async function recordSubproblem(rId,o){ if(!rId||!residents.has(rId)) return {ok:false,message:'no such resident'} ; const id=o.id||('s-'+shortId()); const lines=['# 子问题｜'+(o.title||id),'- 标题: '+(o.title||id),'- ID: '+id,'- 状态: 求解中','- 价值程度: '+cl(o.value!=null?o.value:0.5),'- 动机用途计划: '+(o.motivation||''),'- 依赖: []','','## 陈述',String(o.statement||''),'','## 进度','']; await writeText('Subproblems/'+rId+'/'+id+'.md',lines.join('\n')); logActivity('record',rId+' 子问题 '+id); bumpArtifacts(); return {ok:true,id,file:'Subproblems/'+rId+'/'+id+'.md'} }
     // auto-sync meeting: every meetingKeepEvery new artifacts, convene a general coordination meeting
     function bumpArtifacts(){ artifactCount+=1; markProgress(); if(!meetingState && !verifyState && !pendingMeeting && Number(params.meetingKeepEvery)>0 && artifactCount % Number(params.meetingKeepEvery)===0){ startMeeting('定期同步：分工/进展/是否需要验证','general',null).catch(()=>{}) } }
     function listResidents(){ return Array.from(residents.values()).map(r=>({id:r.rId,direction:r.direction,status:r.status,rounds:r.rounds,contextPct:r.contextPct,insight:r.insight?r.insight.slice(0,80):''})) }
@@ -653,7 +653,7 @@ export function apply(ctx) {
       return (params.residentPersona?params.residentPersona+'\n':'')
         +'Resident researcher '+r.rId+' — CHECKPOINT（团队空闲，请由你们继续自主推进）。当前项目尚未解决（除非你已确认）。团队在等待有人继续：请**继续解决这个问题**——读他人的库对齐、推进某个子问题/引理/方法、尝试一条路线；或向团队发消息（input）、提议任务（propose_task）让大家分工。若你确实认为问题已解决、或已彻底无路可走，才提议开会（propose_meeting）让团队表决/商量、或声明 solved=true。默认立场是：**请推进，而不是停在原地。**\n'
         +'Reply with ONLY a JSON object:\n'
-        +'{"summary":"<what you will do / what you advanced this round>","input":"<optional: a message to the whole team, or \\"\\">","solved":false,"propose_verify":"<id|null>","propose_meeting":"<agenda|null>","propose_task":"<task title|null>","claim_task":"<id|null>","contextPct":40}'
+        +'{"summary":"<what you will do / what you advanced this round>","input":"<optional: a message to the whole team, or \\"\\">","solved":false,"propose_verify":"<id|null>","propose_meeting":"<agenda|null>","propose_task":"<task title|null>","task_desc":"<optional: why this task matters / what it covers|null>","claim_task":"<id|null>","contextPct":40}'
     }
     function clearHeartbeat(){ if(heartbeatDisposer!==null){ try{ heartbeatDisposer() }catch(e){} heartbeatDisposer=null } }
     function armHeartbeat(){
@@ -774,7 +774,13 @@ export function apply(ctx) {
     // ---- resident end handler ----
     async function onResidentEnd(childId, info){
       const r=byChild(childId); if(!r) return
-      busy.delete(r.rId)
+      // A turn that is NOT marked busy is a duplicate/stale end (the same subagent/end delivered twice,
+      // or an end for a turn already settled). Without this guard every side effect below — task
+      // proposal, group relay, verify queueing, meetings.push — would run a SECOND time (the
+      // duplicate-task/duplicate-stop class from test9 reappears whenever a host re-delivers an end).
+      // Every legitimate end corresponds to a busy turn: busy is added at spawn/wake and cleared only
+      // here, on wake failure, on removeMember, or on respawn (whose stale childIds no longer match).
+      if(!busy.delete(r.rId)) return
       // Any resident turn that COMPLETED is real activity for the stall clock (B). Residents frequently
       // write their libraries via direct fs (not the record* tools), so relying only on
       // bumpArtifacts/markProgress would leave lastProgressAt stale and B would fire against an active
@@ -897,12 +903,14 @@ export function apply(ctx) {
     function status(){ return { ok:true, running, phase, autoDone, project:currentProject, residentCount:residents.size,
       residents:listResidents(), busy:[...busy], taskboard:taskboard.length,
       meetingInProgress: !!(meetingState), verifyInProgress: !!(verifyState), pendingVerify: pendingVerify.length?pendingVerify[0].targetId:null, pendingVerifyCount: pendingVerify.length,
+      parkedMeeting: pendingMeeting?pendingMeeting.agenda:null,
       params:['residentCount','compactAfterRounds','compactThreshold','maxParallel','activityTimeoutMs','meetingKeepEvery','verdictMaxRounds','stallAutoMeetingMs','provider','model','residentPersona','toolAllow','toolDeny'].map(k=>k+'='+(Array.isArray(params[k])?params[k].join(','):params[k])).join(', ') } }
     function report(){ return { ok:true, running, phase, autoDone, project:currentProject, problem:problemText,
       residents:listResidents(), taskboard:taskboard.filter(t=>t.status!=='done'),
       meeting: meetingState?{id:meetingState.id, agenda:meetingState.agenda, spoke:Object.keys(meetingState.inputs).length+'/'+residents.size}:null,
       verify: verifyState?{target:verifyState.targetId,stage:verifyState.stage, voted:Object.keys(verifyState.verdicts).length+'/'+residents.size}:null,
       pendingVerify: pendingVerify.length?pendingVerify[0].targetId:null,
+      parkedMeeting: pendingMeeting?pendingMeeting.agenda:null,
       meetings:meetings.length, recentActivity: activityLog.slice(-8) } }
     async function addMember(direction){ const r=newResident(direction||''); await spawnResident(r)
       // Mid-meeting additions must join the meeting's speaking order; otherwise allSpoke (over CURRENT
@@ -914,11 +922,14 @@ export function apply(ctx) {
       return {ok:true,id:r.rId,direction:r.direction} }
     async function removeMember(id){ const r=residents.get(id); if(!r) return {ok:false}; if(r.childId){ try{ subagents.interrupt(r.childId,{kind:'ancestor',agent:rootAgent}) }catch(e){} } residents.delete(id); busy.delete(id); mailboxes.delete(id); wakeKind.delete(id); if(currentResident===id) currentResident=''
       // Reconcile in-progress coordination so a removed member cannot hang consensus or crash a round:
-      // drop its meeting speech / verify verdict / pending-verify proposals if it owned them, and
-      // prune it from the meeting's speaking order so the find() there never selects a ghost.
+      // drop its meeting speech / verify verdict and prune it from the meeting's speaking order so the
+      // find() there never selects a ghost. Its QUEUED verify proposals are deliberately KEPT: a
+      // proposal is a statement about an OBJECT the group can judge on its merits with its CURRENT
+      // members (allVoted recomputes over the live residents), and dropping the queue entry would also
+      // erase the intent of any OTHER member who independently proposed the same target (dedup keeps
+      // only the first entry, which may belong to the removed member).
       if(meetingState){ delete meetingState.inputs[id]; meetingState.order=(meetingState.order||[]).filter(x=>x!==id) }
       if(verifyState){ delete verifyState.verdicts[id] }
-      if(pendingVerify.length) pendingVerify = pendingVerify.filter(p=>p.proposer!==id)
       await saveAll()
       // Re-drive the scheduler right away. If the removed member was the ONLY turn in flight (e.g. the
       // last unspoken meeting speaker / the last unvoted voter, interrupted mid-turn), NO subagent/end
@@ -959,7 +970,7 @@ export function apply(ctx) {
       onResidentEnd, start, resume, status, report, addMember, removeMember, setParams,
       setPause, initAbort, postMessage, startMeeting, saveAll, broadcast, configure, loadSettings,
       currentResident:()=>currentResident,
-      residentIdOf:(agent)=>{ const m=residentOfAgent(agent); return m||currentResident },
+      residentIdOf:(agent)=>{ const m=residentOfAgent(agent); if(m) return m; const c=currentResident; return (c && residents.has(c)) ? c : '' },
       useResident:(id)=>{ currentResident=id },
       publishProgress, recordProposition, recordMethod, recordSubproblem, listResidents, reportContext,
       proposeTask, claimTask, taskDone, listTasks,

package/vibe-math-v4//345/256/236/347/216/260/346/226/271/346/241/210.md

@@ -570,7 +570,7 @@ VibeMath/Projects/<project>/
 >
 > 说明：无法对"某常驻真的拒绝/掉线"的情况达成全体一致时，看门狗会把该对象保留为"未定论/带概率"，这是哲学上期望的诚实结果。`session.json` 仍不持久化 `meetingState`（进行中的会议不跨进程恢复），B 的停滞看门狗在重启后仍会触发，因此重启也能自愈。
 >
-> 全套测试仍全绿：`selfdrive-v4` 21/21、`e2e-v4-fixes` 80/80（含 T20 源卡回写 / T21 终结重入 / T22-23 验证去重 / T24-28 后续审计回归，见 §27）、v3 100/100、v2 business 18/regression 14、multisession 25/25。
+> 全套测试仍全绿：`selfdrive-v4` 21/21、`e2e-v4-fixes` 94/94（T1–T32，后续审计回归见 §27/§28）、v3 100/100、v2 business 18/regression 14、multisession 25/25。
 
 ## 27. 第三轮全面深度审计（npm v2.0.18，牵连对象/逻辑）
 
@@ -587,6 +587,19 @@ VibeMath/Projects/<project>/
 
 > 设计边界复查确认（未改）：① 会议/验证轮次唤醒**不额外受 maxParallel 限制**——共识需要全员发言/投票；每个调度 pass 只唤醒一个未发言/未投票者，由各 end 事件串行驱动（残留的普通轮可能造成短暂并发，但不会无限累积）；② 非一致（平均概率留库）的对象可被后续提议**再次端到端验证**（去重窗口只挡"刚定论为真/假"的对象）——重复提议者是在不知情时提议的，重验一轮不算错；③ 验证中途 `addMember` 的新成员会被要求投票（v2.0.17 注释行为，保持）；④ 已定论 run 的成果随时可在磁盘读取，resume 拒绝只针对"自动复活"。
 
+## 28. 第四轮全面深度审计（npm v2.0.19，对 v2.0.18 修复的牵连检查）
+
+对 v2.0.18 的修复（FIFO 验证队列、会议暂存、暂停门控、writeJson 串行化等）做连带对象复查，发现并修复 4 处（新增回归 T29–T32）：
+
+- **同回合重复 `subagent/end` 会把每个副作用跑两遍**：`onResidentEnd` 对"已结算回合的重复/迟到 end"无守卫——重复投递同一 end（宿主重放、桥接层抖动）会让 propose_task / 群聊转发 / 验证入队 / meetings.push **全部二次执行**（test9 的重复任务/重复 stop 一类问题在"重复 end"场景会原样复发）。修复：回合必须以 busy 标记在册为前提——`if(!busy.delete(r.rId)) return`（每次合法 end 都对应一次 spawn/wake 置位的 busy；busy 仅在本处、唤醒失败、removeMember、重生成时清除，而后者 stale childId 已无法 byChild 命中）。**T29**：同一 tick 内连发两次相同 end → 任务只提议一次、群聊只转发一次。
+- **空/失效 rId 的 resident 写入工具会制造库根散卡**：`currentResident` 被复位为 `''`（如移除最后一名常驻后）时，主机/未知代理调用 `vibe_v4_record_*`/`vibe_v4_publish_progress` 会把路径 `Propos//p-x.md` 塌缩成**库根目录的散乱卡片**（正是 test9 曾出现的 0 字节散卡一类问题）。修复：`residentIdOf` 回退仅当 `currentResident` 仍真实存在时生效，否则返回 `''`；四个库写入函数对空/未知 rId 直接拒绝 `{ok:false}`（主机以 `currentResident` 代写这一既有便利不受影响——自驱动测试仍走该回退）。**T30**。
+- **移除者的已排队验证提议被丢弃 → 提议静默蒸发（含"第二提议者意图被去重吞掉"边角）**：v2.0.12 遗留的"移除提议者则清掉其 pendingVerify"在 FIFO 队列下会把该成员的排队提议**整条删除**；且同目标被两人提议时去重只保留第一人的条目——若第一人被移除，第二人的意图同样蒸发。修复：`removeMember` **不再过滤验证队列**——提议是关于对象的陈述，由**当前成员**按对象价值共识判定（allVoted 本就按实时成员集计算，写回扫描也不依赖提出者存续）。**T31**：r-3 提议 p-x 后在 p-y 验证中被移除 → p-x 仍由剩余成员完整验证到 Verified/。
+- **暂存的会议对用户不可见 / 普通轮提议任务永远没有描述**：`status/report` 新增 `parkedMeeting`（暂存会议议程），便于确认"会议没丢、在排队"（**T32** 同时钉住"brainstorm 期开会 → 暂存 → bootstrap 后真正召开"，不再被看门狗误杀）；普通轮/心跳提示词模板补上 `task_desc` 字段（此前普通轮提议任务时描述恒为空，只有会议轮能带描述）。
+
+> 边界复查（未改，记录）：brainstorm 阶段若某常驻首轮 end 永不触发，run 停在 brainstorm（brainstorm 无看门狗、无心跳可唤醒"已空闲"者推进）——真实 DSH 的轮次必然以 completed/error/timeout 结束并触发 end，纯属框架外故障，不做可能"静默丢掉慢成员声音"的自动跳过；`Shared/taskboard.md`（人类视图）由调用时快照生成、非逐文件串行，最坏情况短暂滞后于 `State/taskboard.json`（权威源），下次任务动作即自愈。
+>
+> 全套测试仍全绿：`selfdrive-v4` 21/21、`e2e-v4-fixes` 94/94（T1–T32）、v3 100/100、v2 business 18/regression 14、multisession 25/25、selfdrive-v3 0 异常。
+