agentlas 1.0.35 → 1.0.37

package/CHANGELOG.md

@@ -1,5 +1,22 @@
 # Changelog
 
+## 1.0.37 — 2026-08-10
+
+Agentlas One can now carry its owner-bound identity and curated memory tickets
+across Terminal sessions without turning the agent into the owner of a project.
+
+- When One is explicitly enabled, Terminal loads its bounded directive at the
+  per-turn system boundary and forwards only `agent_repo` and `user_identity`
+  memory candidates to One's existing ledger. Project memory stays with the
+  project, and a missing or disabled One workspace remains a no-op.
+- The Memory Events parser accepts both the canonical ticket envelope and the
+  legacy array form, so valid candidates are no longer silently discarded.
+- Global `--json` output now reaches doctor, list, roles, Cloud restore, and
+  upload commands consistently; machine-readable output is no longer prefixed
+  by human status text.
+- The built-in architecture projection includes Agentlas One as a hidden
+  orchestrator rather than a top-level project or user-facing worker.
+
 ## 1.0.34 — 2026-08-06
 
 Telegram, standalone. The terminal can now connect a Telegram bot with no desktop app.

package/engine/architecture.data.json

@@ -1,5 +1,5 @@
 {
-  "version": "1.6.0",
+  "version": "1.7.0",
   "emitterBlock": "## Memory (Agentlas curated memory)\n\nAt the end of EVERY completed normal reply, emit exactly one hidden Memory Events\nenvelope. The runtime removes it before display. This envelope is the per-turn receipt:\nalways include a compact safe turn_summary, and use an empty candidates array when\nnothing durable was learned. Do not skip the envelope.\n\nRules:\n- Never include secrets, credentials, API keys, raw logs, or full transcripts.\n- Real credential values may live only in local project .env/.env.local,\n  ignored signing/ or credentials/ files, or a local keychain/vault. Memory\n  Events may mention env names and local relative paths only.\n- For deploy, release, store, billing, auth, API, or cloud work, first read the\n  project's .agentlas/local-credentials.map.json and the top\n  \"Local Credential Index\" section of .agentlas/project-soul-memory.md\n  before saying a credential is missing.\n- One candidate per durable item. Keep \"content\" to one or two sentences.\n- \"memory_kind\": fact | decision | preference | risk | procedure | hypothesis | evidence | deprecation | conflict\n- \"suggested_scope\": user_identity | team_memory | project (this folder) | agent_repo | session (temporary) | discard\n- Use user_identity for a stable operator preference or personal fact (their name, role, language, tone,\n  how they want you to behave) — these must outlive any one project. The curator only files user_identity\n  when you label it so with \"confidence\": \"high\"; it never promotes into that scope, so a preference emitted\n  at lower confidence is demoted to a throwaway session note.\n- \"agent_team\" is accepted only as a legacy alias for team_memory.\n- Add \"request_context\" when it improves future recall: user_intent, trigger_terms,\n  cwd_at_request, target_project, target_path, cross_context, outcome.\n- Never put the raw user prompt or transcript in request_context.\n- Suggest a scope; the separate Memory Curator decides the final destination.\n- turn_summary is one value-free sentence about the completed outcome. It is not the\n  user prompt, a transcript, raw log, secret, or absolute local path.\n\nFormat (always emit, including an empty candidates array):\n\n## Memory Events\n```json\n{\n  \"schema_version\": \"agentlas.memory-ticket.v1\",\n  \"turn_summary\": \"Completed outcome in one safe sentence.\",\n  \"candidates\": [\n    {\n      \"memory_kind\": \"decision\",\n      \"content\": \"...\",\n      \"suggested_scope\": \"project\",\n      \"confidence\": \"high\",\n      \"sensitivity\": \"internal\",\n      \"evidence_refs\": [],\n      \"request_context\": {\n        \"user_intent\": \"...\",\n        \"trigger_terms\": [\"...\"],\n        \"cwd_at_request\": null,\n        \"target_project\": null,\n        \"target_path\": null,\n        \"cross_context\": false,\n        \"outcome\": \"...\"\n      }\n    }\n  ]\n}\n```",
   "eventsHeading": "## Memory Events",
   "memoryDir": ".agentlas",
@@ -138,6 +138,18 @@
       "visibility": "background",
       "tone": "amber",
       "systemPrompt": "# Task Bias Curator (Agentlas built-in)\n\nYou reduce TASK BIAS in multi-surface projects — the tendency to keep working on\nsurfaces that are recent, salient, or easy to measure while other surfaces stay\nuninspected. You are a SECOND-ORDER control role: you adjust the rules of work\nallocation and evidence review; you do not implement product work yourself, and you\ncannot mark a node \"complete\".\n\n## External state: the AI Sitemap\nThe project's shared external state lives in .agentlas/sitemap.json. Each\nnode carries: node_id, kind, status (unknown|todo|in_progress|blocked|validated|revalidate),\ncompletion_score (0..1, evidence-backed), risk_level, last_modified, last_tested,\ndependencies, acceptance_checks, evidence, provisional.\n\n## What you do\n1. Read/maintain the sitemap. Create provisional nodes for newly discovered surfaces.\n2. Choose the next bounded task from a VISIBLE priority policy, not recent chat context:\n   prioritize high risk, low completion_score, stale last_tested, and blocking dependencies.\n3. Audit for bias: which surfaces are over-worked vs never inspected? Name them.\n4. Audit validation: flag completion claims without evidence or with weak evidence;\n   require revalidation and name the missing evidence.\n5. Produce a compact, reversible curator decision record. Escalate mission-level changes\n   to the user.\n\n## Boundaries\nCannot mark a node complete. Cannot erase evidence (only supersede it with a logged\ndecision). Cannot expand the project mission without explicit user approval.\n\nKeep outputs small: a policy/priority recommendation, a revalidation request, a\nsitemap update proposal, or a provisional-node decision."
+    },
+    {
+      "id": "builtin-agentlas-orchestrator",
+      "slug": "agentlas-one",
+      "name": "Agentlas One",
+      "nameEn": "Agentlas One",
+      "tagline": "오너 전속 개인 에이전트 — 전문가·도구·기억을 필요할 때만 꺼내 쓴다",
+      "taglineEn": "Owner-bound personal agent that pulls in specialists, tools, and memory on demand",
+      "role": "orchestrator",
+      "visibility": "background",
+      "tone": "green",
+      "systemPrompt": "# Agentlas One (Agentlas built-in)\n\n너는 채팅 어시스턴트가 아니라 오너 전속 개인 에이전트다. 세션·프로젝트·런타임을 넘어\n같은 정체성을 유지한다. 자기소개를 반복하지 않는다.\n\n## 일하는 법\n- 전문가가 필요하면 새로 고르기 전에 이미 묶인 로스터를 재사용한다. 채용은 가산이다.\n- \"못 한다\"고 말하기 전에 보유 수단(로컬 에이전트·오너 클라우드·Hub·플러그인)을 실제로 조회한다.\n  호출하지 않은 도구를 호출한 척하지 않는다.\n- 되돌리기 어려운 일(파일 삭제·발송·결제·공개) 직전에는 기억보다 실측을 우선한다.\n  확인이 안 되면 진행하지 말고 오너에게 묻는다.\n- 기억이 서로 모순되면 하나를 골라 단정하지 말고 충돌을 그대로 말한다.\n- 막히면 멈추지 말고 남은 수단을 순서대로 시도하고, 다 막히면 어디서 왜 막혔는지 기계 근거로 한 줄.\n\n## 기억\ndurable 기억을 직접 쓰지 않는다. 관찰은 Memory Events 로 내고 런타임이 티켓으로 포장해\n큐레이터에 넘긴다. 증거 없는 fact/decision/procedure 는 hypothesis 다.\n\n## 경계\n허브·클라우드에 업로드되지 않는다. 오너의 원시 기억·자격증명·전사를 외부로 내보내지 않는다."
     }
   ]
 }

package/engine/commands/cloud.cjs

@@ -7,7 +7,10 @@ const { runCloud } = require("../cloud-assets/commands.cjs");
 
 async function run(ctx, args) {
   try {
-    return await runCloud(ctx, args);
+    const commandArgs = ctx.output?.format === "json" && !args.includes("--json")
+      ? [...args, "--json"]
+      : args;
+    return await runCloud(ctx, commandArgs);
   } catch (e) {
     ctx.err(String((e && e.message) || e));
     return 1;

package/engine/commands/doctor.cjs

@@ -27,7 +27,7 @@ function roleDetail(selection, role, en) {
 function run(ctx, args = []) {
   const en = ctx.lang === "en";
   // clig.dev: 스크립트 소비자를 위한 기계 계약. 사람용 줄과 같은 사실만 담는다.
-  if (args.includes("--json")) {
+  if (ctx.output?.format === "json" || args.includes("--json")) {
     const db = ctx.db();
     const clis = listAvailableCliRuntimes().map((c) => ({ kind: c.kind, path: c.path, authEvidence: runtimeAuthEvidence(c.kind).status }));
     const active = activeRuntimeRow(db);

package/engine/commands/list.cjs

@@ -36,7 +36,7 @@ function run(ctx, args = []) {
 
   // clig.dev: 스크립트 소비자는 사람용 표를 파싱하게 두지 말 것 — --json 은
   // 사람용 출력과 같은 사실을 기계 계약으로 준다.
-  if (args.includes("--json")) {
+  if (ctx.output?.format === "json" || args.includes("--json")) {
     const orchestrator = resolvedModelRole(db, "orchestrator");
     const worker = resolvedModelRole(db, "worker");
     ctx.out(JSON.stringify({ agents, firms, modelRoles: { orchestrator, worker } }, null, 2));

package/engine/commands/roles.cjs

@@ -38,7 +38,7 @@ function fmt(selection, en) {
 function show(ctx, args = []) {
   const en = ctx.lang === "en";
   const db = ctx.db();
-  if (args.includes("--json")) {
+  if (ctx.output?.format === "json" || args.includes("--json")) {
     ctx.out(JSON.stringify({
       orchestrator: resolvedModelRole(db, "orchestrator"),
       worker: resolvedModelRole(db, "worker"),
@@ -177,6 +177,7 @@ function set(ctx, args) {
 
 function run(ctx, args = []) {
   const en = ctx.lang === "en";
+  if (ctx.output?.format === "json" && !args.includes("--json")) args = [...args, "--json"];
   const [sub, ...rest] = args;
   if (!sub || sub === "show" || sub === "list" || sub === "--json") return show(ctx, sub === "--json" ? ["--json"] : rest);
   if (sub === "set") return set(ctx, rest);

package/engine/commands/upload.cjs

@@ -8,7 +8,10 @@ const { runUpload } = require("../cloud-assets/commands.cjs");
 
 async function run(ctx, args) {
   try {
-    return await runUpload(ctx, args);
+    const commandArgs = ctx.output?.format === "json" && !args.includes("--json")
+      ? [...args, "--json"]
+      : args;
+    return await runUpload(ctx, commandArgs);
   } catch (e) {
     ctx.err(String((e && e.message) || e));
     return 1;

package/engine/memory-cli/curate.cjs

@@ -96,7 +96,18 @@ function parseMemoryEventsCli(text) {
   const after = text.slice(idx + heading.length);
   const fence = after.match(/```(?:json)?\s*([\s\S]*?)```/);
   let events = [];
-  if (fence) { try { const d = JSON.parse(fence[1].trim()); if (Array.isArray(d)) events = d; } catch { /* ignore */ } }
+  // 봉투 두 형태를 모두 받는다. Desktop `electron/memory/events.ts:128,156` 과 같은 계약이다.
+  //  · 정본: {"schema_version":"agentlas.memory-ticket.v1","candidates":[...]}  ← 이미터가 지시하는 형태
+  //  · 레거시: [...]  (최상위 배열)
+  // ★배열만 받던 시절 정본 봉투는 **조용히 버려졌다**(실측: 터미널이 후보 5건을 냈는데 파서 결과 0건).
+  // 형태가 낯설면 버리지 말고, 최소한 왜 못 읽었는지 남길 수 있게 둘 다 통과시킨다.
+  if (fence) {
+    try {
+      const d = JSON.parse(fence[1].trim());
+      if (Array.isArray(d)) events = d;
+      else if (d && typeof d === "object" && Array.isArray(d.candidates)) events = d.candidates;
+    } catch { /* ignore */ }
+  }
   let cut = text.length;
   if (fence && fence.index != null) cut = idx + heading.length + fence.index + fence[0].length;
   const before = text.slice(0, idx).replace(/<!--\s*$/u, "");

package/engine/sessions/apply-fences.cjs

@@ -18,6 +18,75 @@ const memoryCurate = require("../memory-cli/curate.cjs");
 const automationStore = require("../automation/store.cjs");
 const schedule = require("../automation/schedule.cjs");
 
+/** One 이 가져가는 스코프. 프로젝트 스코프는 프로젝트에 남는다. */
+const ONE_SCOPES = new Set(["agent_repo", "user_identity"]);
+
+/**
+ * Agentlas One 서랍(`~/.agentlas/one/.agentlas/memory-tickets.jsonl`)으로 후보를 넘긴다.
+ *
+ * One 은 프로젝트를 넘나드는 정체성이라 agent_repo/user_identity 만 가져간다.
+ * One 이 꺼져 있거나 서랍이 없으면 아무것도 하지 않는다 — 없는 폴더를 만들지 않는다.
+ * 실패해도 턴을 죽이지 않는다(펜스 적용 계약과 동일).
+ */
+function forwardToOne(events) {
+  try {
+    const fs = require("node:fs");
+    const os = require("node:os");
+    const path = require("node:path");
+    const root = process.env.AGENTLAS_ONE_DIR || path.join(os.homedir(), ".agentlas", "one");
+    const state = JSON.parse(fs.readFileSync(path.join(root, "state.json"), "utf8"));
+    if (!state || state.on !== true) return 0;
+    const ledger = path.join(root, ".agentlas", "memory-tickets.jsonl");
+    if (!fs.existsSync(ledger)) return 0;
+
+    // 이미 올라온 내용은 다시 넣지 않는다(같은 계약을 One 쪽 emit_ticket 도 쓴다).
+    const seen = new Set();
+    for (const line of fs.readFileSync(ledger, "utf8").split("\n")) {
+      if (!line.trim()) continue;
+      try {
+        const row = JSON.parse(line);
+        const text = String((row.candidate || {}).content || "");
+        if (text) seen.add(text.trim().toLowerCase().replace(/\s+/g, " "));
+      } catch { /* 깨진 줄은 건너뛴다 */ }
+    }
+
+    let written = 0;
+    for (const raw of events) {
+      for (const candidate of (raw && Array.isArray(raw.candidates) ? raw.candidates : [raw])) {
+        if (!candidate || typeof candidate !== "object") continue;
+        const scope = String(candidate.suggested_scope || candidate.scope || "");
+        if (!ONE_SCOPES.has(scope)) continue;
+        const content = String(candidate.content || "").trim();
+        if (!content) continue;
+        const key = content.toLowerCase().replace(/\s+/g, " ");
+        if (seen.has(key)) continue;
+        seen.add(key);
+        const evidence = Array.isArray(candidate.evidence) ? candidate.evidence.slice(0, 8) : [];
+        fs.appendFileSync(ledger, JSON.stringify({
+          schemaVersion: "agentlas.one-workspace.v1",
+          ticketId: `one-tkt-${Date.now()}-${written}`,
+          agentId: "builtin-agentlas-one",
+          turnKey: "",
+          source: "terminal-memory-events",
+          state: "queued",
+          candidate: {
+            type: String(candidate.memory_kind || candidate.type || "hypothesis"),
+            scope,
+            content: content.slice(0, 600),
+            evidence,
+          },
+          downgraded: false,
+          createdAt: new Date().toISOString().replace(/\.\d+Z$/, "Z"),
+        }) + "\n", "utf8");
+        written += 1;
+      }
+    }
+    return written;
+  } catch {
+    return 0;
+  }
+}
+
 /**
  * @param {import('./session.cjs').Session} session 방금 턴을 끝낸 세션
  * @param {object} parsed parseReplyFences 결과
@@ -69,6 +138,10 @@ function applyReplyFences(session, parsed, opts = {}) {
       written: ctx.curatedMemories.length,
       permission: session.permission,
     };
+    // Agentlas One 이 켜져 있으면 에이전트 스코프 후보를 One 서랍에도 티켓으로 넘긴다.
+    // 프로젝트 스코프는 여기 남기고 옮기지 않는다 — One 은 프로젝트를 넘나드는 정체성이라
+    // agent_repo/user_identity 만 One 의 것이다(기획 2.2 스코프 경계).
+    receipts.memory.one = forwardToOne(parsed.memoryEvents);
     // read 권한 턴 = durable 쓰기 0 — 영수증 이벤트만 남는다.
     record({ type: "memory-curated", ...receipts.memory });
   }
@@ -185,4 +258,5 @@ function applyReplyFences(session, parsed, opts = {}) {
   return receipts;
 }
 
-module.exports = { applyReplyFences };
+// forwardToOne 은 One 서랍 전달의 유일한 지점이라 계약 테스트가 직접 잴 수 있게 함께 노출한다.
+module.exports = { applyReplyFences, forwardToOne };

package/engine/sessions/prompt.cjs

@@ -180,6 +180,29 @@ function cliMemoryContext(db, projectPath, agentId = null, task = "") {
  * 최종 시스템 프롬프트 조립. ctx = { lang, projectPath, agentId, turnId, permission }.
  * withEmitter=false 는 이미터/리마인더 없이(캡처·판정 등 내부 턴용).
  */
+/**
+ * Agentlas One 지시문. 켜져 있을 때만, 그리고 정본 파일이 실재할 때만 싣는다.
+ * 상태 파일이 없으면 조용히 빈 문자열 — 꺼진 One 의 지시문을 흘리지 않는다.
+ */
+function loadOneDirective() {
+  try {
+    const root = process.env.AGENTLAS_ONE_DIR
+      || require("node:path").join(require("node:os").homedir(), ".agentlas", "one");
+    const fs = require("node:fs");
+    const path = require("node:path");
+    const state = JSON.parse(fs.readFileSync(path.join(root, "state.json"), "utf8"));
+    if (!state || state.on !== true) return "";
+    const text = fs.readFileSync(path.join(root, "directive.md"), "utf8").trim();
+    // 상한을 둔다 — 지시문이 남의 토큰 예산을 잠식하면 안 된다(메모리 예산과 같은 규칙).
+    return text.length > ONE_DIRECTIVE_MAX_CHARS ? text.slice(0, ONE_DIRECTIVE_MAX_CHARS) : text;
+  } catch {
+    return "";
+  }
+}
+
+/** [튜닝값, 근거 없음] — 현재 정본 지시문이 약 3.7KB 라 두 배 여유를 둔다. */
+const ONE_DIRECTIVE_MAX_CHARS = 8000;
+
 function augmentSystem(db, baseSystem, ctx, withEmitter, request = "") {
   const arch = loadArch();
   let sys = baseSystem || "";
@@ -188,6 +211,10 @@ function augmentSystem(db, baseSystem, ctx, withEmitter, request = "") {
   sys = responseDirective(lang) + (sys ? "\n\n" + sys : "");
   const connectionSkill = loadGlobalConnectionSkill();
   if (connectionSkill) sys += "\n\n" + connectionSkill;
+  // Agentlas One 이 켜져 있으면 그 지시문을 싣는다. R4 기준 터미널의 "매 턴 주입 지점"이 여기다.
+  // 정본은 `~/.agentlas/one/directive.md` 하나 — CLAUDE.md/AGENTS.md 의 마커 블록은 그 사본이다.
+  const oneDirective = loadOneDirective();
+  if (oneDirective) sys += "\n\n" + oneDirective;
   const mem = cliMemoryContext(db, ctx && ctx.projectPath, ctx && ctx.agentId, request);
   if (mem) sys += "\n\n" + mem;
   // 도구 접근 고지 — 터미널에는 이게 아예 없었다. 도구가 붙지 않은 턴에서 CLI는 아무

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agentlas",
-  "version": "1.0.35",
+  "version": "1.0.37",
   "description": "Agentlas project terminal — run project controllers and task-scoped agent teams from the terminal. Standalone: no desktop app process required.",
   "bin": {
     "agentlas": "bin/agentlas.cjs"