fix: unify agentrun session send cli

.agents/skills/unidesk-code-queue/SKILL.md

@@ -1,11 +1,11 @@
 ---
 name: unidesk-code-queue
-description: UniDesk AgentRun-backed Code Queue CLI — Skill(cli-spec)。legacy `codex` 子命令只保留历史只读和残留停止；新任务提交、Aipod/Artificer 派单、steer/send、events/logs/result、ack/cancel、dispatch、run/command/runner 状态 drill-down 和 HWLAB Code Agent/CaseRun follow-up 必须使用 `agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send` 资源原语。用户提到 codex、Code Queue、submit、steer、resume、tasks、unread、code-queue、aipod、Artificer、HWLAB Code Agent 时使用。
+description: UniDesk AgentRun-backed Code Queue CLI — Skill(cli-spec)。legacy `codex` 子命令只保留历史只读和残留停止；新任务提交、Aipod/Artificer 派单、session follow-up、events/logs/result、ack/cancel、dispatch、run/command/runner 状态 drill-down 和 HWLAB Code Agent/CaseRun follow-up 必须使用 `agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send` 资源原语。用户提到 codex、Code Queue、submit、send、resume、tasks、unread、code-queue、aipod、Artificer、HWLAB Code Agent 时使用。
 ---
 
 # UniDesk Code Queue / AgentRun CLI
 
-旧 Code Queue 已冻结新任务和写入口。`bun scripts/cli.ts codex ...` 现在只作为历史归档、只读排障和残留任务停止入口；新的指挥官派单、Aipod/Artificer 执行、events/logs/result、ack/cancel、dispatch、steer/send 必须走 AgentRun 资源原语，并按 cli-spec 渐进披露。UniDesk 是 render-only client：默认输出是低噪声 human 表格/摘要，脚本读取显式使用 `-o json|yaml` 的稳定客户端 schema，`--raw` 只用于查看直连 AgentRun REST envelope。
+旧 Code Queue 已冻结新任务和写入口。`bun scripts/cli.ts codex ...` 现在只作为历史归档、只读排障和残留任务停止入口；新的指挥官派单、Aipod/Artificer 执行、events/logs/result、ack/cancel、dispatch、session follow-up 必须走 AgentRun 资源原语，并按 cli-spec 渐进披露。UniDesk 是 render-only client：默认输出是低噪声 human 表格/摘要，脚本读取显式使用 `-o json|yaml` 的稳定客户端 schema，`--raw` 只用于查看直连 AgentRun REST envelope。
 
 **固定入口前缀**: `cd /root/unidesk && bun scripts/cli.ts agentrun ...`
 
@@ -55,11 +55,10 @@ bun scripts/cli.ts agentrun logs session/<sessionId> --tail 100
 bun scripts/cli.ts agentrun ack session/<sessionId>
 bun scripts/cli.ts agentrun dispatch task/<taskId>
 bun scripts/cli.ts agentrun send session/<sessionId> --aipod Artificer --prompt-stdin
-bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin
 bun scripts/cli.ts agentrun cancel session/<sessionId> --reason <text> --dry-run
 ```
 
-日常 task manifest 优先使用 YAML heredoc：`agentrun apply -f -`；单 prompt 派单优先 `agentrun create task --aipod Artificer --prompt-stdin`。UniDesk 客户端按 `config/agentrun.yaml` 直连 AgentRun REST API，不经过 HWLAB runtime、SSH official CLI 或旧 bridge wrapper；`--json-file`、`--prompt-file` 和 `--runner-json-file` 只是客户端输入来源，用于已审阅且可复用的受控文件。它不是旧 Code Queue adapter，不双写，也不迁移旧历史。
+日常 task manifest 优先使用 YAML heredoc：`agentrun apply -f -`；单 prompt 派单优先 `agentrun create task --aipod Artificer --prompt-stdin`；同 session 续跑只使用 `agentrun send session/<sessionId>`。UniDesk 客户端按 `config/agentrun.yaml` 直连 AgentRun REST API，不经过 HWLAB runtime、SSH official CLI 或旧 bridge wrapper；`send` 是唯一用户级 session follow-up 写入口，服务端按 durable session/run/command 状态自动决定内部 `steer` 或新 `turn`，旧 CLI `turn/steer` 路径不保留兼容。`--json-file`、`--prompt-file` 和 `--runner-json-file` 只是客户端输入来源，用于已审阅且可复用的受控文件。它不是旧 Code Queue adapter，不双写，也不迁移旧历史。
 
 `AipodSpec` 是 AgentRun v0.1 的声明式 agent 装配：模型 profile、gitbundle、skills/tools、SecretRef 和 tool credential 都从 YAML 规格渲染。`Artificer` 默认用于 UniDesk 分布式开发任务，使用 `sub2api` provider、`gpt-5.5`、`reasoningEffort=xhigh`，并通过 SecretRef 注入 GitHub PR token、GitHub SSH 和 UniDesk SSH 透传能力。更新规格时使用 `agentrun aipod-specs apply --yaml-stdin --dry-run` 先看计划，确认后再去掉 `--dry-run`；不得把 API key、SSH key 或 token 写入 prompt、payload、YAML 或 issue。
 
@@ -74,14 +73,14 @@ AgentRun queue 生命周期不是一个单独的 `queue lifecycle` 命令，而
 3. Run 级状态用 `events run/<runId>` 和 `result run/<runId> --command <commandId>`，判断 terminalClassification、failureKind、provider interruption、timeoutBudget 和 recoveryActions。
 4. Command 级状态用 `describe command/<commandId> --run <runId>` 和 `result command/<commandId> --run <runId>`，确认 command state、ack、terminal status 和结果摘要。
 5. Runner job 只读状态用 `describe runnerjob/<runnerJobId> --run <runId>`，确认 env image reuse、jobName、namespace、phase、exitCode、retention 和 `valuesPrinted=false`。不要为了这些字段手动调用 `trans G14:k3s kubectl ...`。
-6. Session trace/output 只在 `describe task` 或 result 里有实际 `sessionId` 时使用 `logs|ack|steer|send|cancel session/<sessionId>`；`sessionRef=null` 时不要猜 session 命令。
+6. Session trace/output 只在 `describe task` 或 result 里有实际 `sessionId` 时使用 `logs|ack|send|cancel session/<sessionId>`；`sessionRef=null` 时不要猜 session 命令。
 7. 已创建但尚未运行的 task 使用 `dispatch task/<taskId>` 派发，不再退回旧 bridge `queue dispatch`。
 
 默认视图必须低噪声且不是 JSON envelope，`-o json|yaml` 才输出稳定机器结构，`--raw` 才保留直连 AgentRun REST envelope；命令返回里的下一步应优先是 `bun scripts/cli.ts agentrun ...` 资源原语，不得把人工 k8s 查询作为日常下一步。
 
 ## HWLAB Code Agent 入口整合
 
-HWLAB Code Agent / CaseRun follow-up 的日常派单也归入 AgentRun 资源原语：新任务用 `create task --aipod Artificer` 或包含 HWLAB gitbundle 的 `apply -f -`；运行中纠偏用 `steer session/<sessionId>` 或 `send session/<sessionId> --aipod Artificer`。需要验证 HWLAB Web/Cloud API 原入口时，仍按 `$hwlab-code-agent` 使用 G14 `/root/hwlab-v02` 的 `hwlab-cli client agent ...` 拉取同一 trace/result/inspect；不要回到旧 `codex submit/resume/steer`。
+HWLAB Code Agent / CaseRun follow-up 的日常派单也归入 AgentRun 资源原语：新任务用 `create task --aipod Artificer` 或包含 HWLAB gitbundle 的 `apply -f -`；运行中纠偏用 `send session/<sessionId> --aipod Artificer`。需要验证 HWLAB Web/Cloud API 原入口时，仍按 `$hwlab-code-agent` 使用 G14 `/root/hwlab-v02` 的 `hwlab-cli client agent ...` 拉取同一 trace/result/inspect；不要回到旧 `codex submit/resume/steer`。
 
 ---
 

docs/reference/cli.md

@@ -92,10 +92,10 @@ CI/CD、GitOps、rollout、artifact 发布、PR 合并后的 runtime lane 滚动
 - `ci install|install-status|status|run|publish-backend-core|publish-user-service|run-dev-e2e|logs` 管理 D601 原生 k3s 上的 Tekton CI。`install` 默认创建 `.state/jobs` 异步 job 并立即返回，`install-status <jobId|latest>` 读取阶段化 progress 和 bounded log tail；只有现场同步调试才显式加 `--wait`。`run` 手动创建每 commit 检查和 Code Queue 只读性能门禁；`publish-backend-core` 与 `publish-user-service` 从 pushed Git commit 构建并发布 `127.0.0.1:5000/unidesk/<service>:<commit>` commit-pinned artifacts，输出 `artifactSummary`（含 `serviceId`、`sourceCommit`、`sourceRepo`、`dockerfile`、`imageRef`、`tag`、`digest`、`digestRef`），但不部署生产；`run-dev-e2e` 的 Git 控制 runner、短 launcher、host fetch 边界、临时 smoke namespace 和 no-CD 规则只在 `docs/reference/dev-ci-runner.md` 定义；Tekton CI 通用规则见 `docs/reference/ci.md`。
 - `schedule list|get|runs|run|retry-run|delete|upsert-pgdata-backup` 管理 backend-core 定时任务和运行历史。`schedule list`、`schedule get`、`schedule runs --limit N` 和 `schedule runs <scheduleId> --limit N` 是只读观察入口；`schedule run`、`schedule retry-run`、`schedule delete` 和 `schedule upsert-pgdata-backup` 会触发运行或写入配置，生产恢复时必须有明确授权。`schedule runs --limit N` 是全局历史视图，返回 `scope=global` 和 `scheduleId=null`；`schedule runs <scheduleId> --limit N` 是指定 schedule 历史视图，返回 `scope=schedule` 和对应 `scheduleId`。CLI 必须拒绝 `schedule runs 50` 这类纯数字位置参数，并提示使用 `schedule runs --limit 50`，避免把空数组误判成“没有历史 run”。`schedule run <id> --wait-ms N` 触发同一 schedule，并且即使 wait 超时也必须返回 `newRunId` 和 `observeCommand`；`schedule retry-run <failedRunId>` 只接受 failed run，从原 run 反查 `scheduleId` 后重触发同一 schedule，并输出 `originalRunId`、`scheduleId`、`newRunId` 和 `observeCommand`。当 backend-core 目标容器缺失或只观察到 verify-only 容器时，schedule/microservice 命令必须以非零退出并返回 `failureKind=target-stack-not-running`、`runnerDisposition=infra-blocked`、`readOnlyCommands` 和 `authorizationRequiredForRecovery`，不得把 Docker 的 `No such container` 当成成功的空历史。
 - `codex deploy <commitId>` 是旧 Code Queue 兼容部署入口，已禁用以防止维护通道直连 D601 部署 Code Queue；当前 dev 自动化只做 `ci run-dev-e2e` smoke，不提供 Code Queue CD，详细规则见 `docs/reference/codex-deploy.md`。
-- `agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send` 是当前指挥官新任务和 AgentRun session 控制入口。UniDesk CLI 是 render-only client：客户端保留 k8s 风格命令解析、human 表格、生命周期摘要、下一步命令、分页、`-o json|yaml` 稳定客户端 schema 和错误展示；AgentRun 服务端只提供稳定 RESTful API、鉴权和业务事实，不承载 UniDesk CLI 渲染。日常查看用 `get tasks --queue commander`、`describe task/<taskId>`、`events run/<runId>`、`logs session/<sessionId>`、`result run/<runId> --command <commandId>`；日常写入用 `create task --aipod Artificer --prompt-stdin`、`apply -f -`、`dispatch task/<taskId>`、`steer/send session/<sessionId>`、`ack/cancel task|session/<id>`。兼容 group `queue|runs|commands|runner|sessions|aipod-specs` 也走同一 direct HTTP transport，`--raw` 只披露直连 AgentRun REST envelope。
+- `agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send` 是当前指挥官新任务和 AgentRun session 控制入口。UniDesk CLI 是 render-only client：客户端保留 k8s 风格命令解析、human 表格、生命周期摘要、下一步命令、分页、`-o json|yaml` 稳定客户端 schema 和错误展示；AgentRun 服务端只提供稳定 RESTful API、鉴权和业务事实，不承载 UniDesk CLI 渲染。日常查看用 `get tasks --queue commander`、`describe task/<taskId>`、`events run/<runId>`、`logs session/<sessionId>`、`result run/<runId> --command <commandId>`；日常写入用 `create task --aipod Artificer --prompt-stdin`、`apply -f -`、`dispatch task/<taskId>`、`send session/<sessionId>`、`ack/cancel task|session/<id>`。用户级 CLI 取消 `turn` 和 `steer` 路径；`send session/<sessionId>` 是唯一 session follow-up 写入口，AgentRun 服务端按 durable session/run/command 状态自动决定内部 `steer` 或新 `turn`，dry-run 必须真实返回这个 decision 且不写状态。兼容 group `queue|runs|commands|runner|sessions|aipod-specs` 也走同一 direct HTTP transport，`--raw` 只披露直连 AgentRun REST envelope。
 - `agentrun` 资源原语的默认 transport 是直连 AgentRun REST API，配置来源是 UniDesk 自有 YAML `config/agentrun.yaml`。鉴权可以复用 `HWLAB_API_KEY` 的环境变量/固定文件发现风格，但不得依赖 HWLAB runtime、HWLAB backend-core、HWLAB frontend 代理或 SSH official CLI；多一层转发会增加故障面，不能作为正式路径。`agentrun control-plane ...` 和 `git-mirror ...` 仍属于 G14 source/runtime 运维控制路径，可以继续使用 UniDesk SSH capture bridge；这些控制面路径不得反向成为 queue/session 资源原语的默认 transport。
 - `agentrun control-plane expose --dry-run|--confirm` 按 `config/agentrun.yaml` 维护 AgentRun 公网 HTTPS 入口，模式与 Sub2API 暴露一致：G14 AgentRun runtime 通过 frpc 出到 master `127.0.0.1:<remotePort>`，master Caddy 提供 `https://agentrun.74-48-78-17.nip.io/`。该命令只补 master `frps` allow port 和 Caddy vhost；G14 frpc Deployment/ConfigMap 必须由 AgentRun `deploy/deploy.json` + GitOps render 管理，不能在 UniDesk 侧手写 Kubernetes manifest。
-- `codex submit/enqueue`、`codex steer`、`codex resume`、`codex queue create`、`codex queue merge`、`codex move`、旧 Web 提交表单、旧队列管理和旧 workdir 管理是冻结的 legacy Code Queue 写入口。CLI 必须返回 `ok=false`、`frozen=true`、`degradedReason=legacy-code-queue-frozen` 和 AgentRun 替代命令；服务端旧 API 写入口必须返回 410。新任务、steer/send、events/logs/result、ack 和 cancel 走 AgentRun 资源原语。
+- `codex submit/enqueue`、`codex steer`、`codex resume`、`codex queue create`、`codex queue merge`、`codex move`、旧 Web 提交表单、旧队列管理和旧 workdir 管理是冻结的 legacy Code Queue 写入口。CLI 必须返回 `ok=false`、`frozen=true`、`degradedReason=legacy-code-queue-frozen` 和 AgentRun 替代命令；服务端旧 API 写入口必须返回 410。新任务、session follow-up、events/logs/result、ack 和 cancel 走 AgentRun 资源原语，其中 session follow-up 只用 `agentrun send session/<sessionId>`。
 - 旧 Code Queue 只保留历史归档、只读排障和残留任务停止。`codex task/tasks/output/read/unread/queues` 继续通过 backend-core 私有代理读取旧 PostgreSQL 历史；`codex interrupt|cancel <taskId>` 只用于停止旧运行面残留任务。旧 `steer-confirm` 只作为历史 trace confirmation 查询，不是新任务控制入口。
 - `codex pr-preflight [--remote] [--push-dry-run --push-dry-run-ref refs/heads/probe/<name>] [--pr-create-dry-run --pr-create-dry-run-head <head>] [--issue N] [--full|--raw]` 通过稳定 `code-queue` proxy 请求 D601 scheduler `/api/runtime-preflight`，用于 PR 型派单 admission。默认输出是紧凑 commander 视图，显式分出 `schedulerPreflight` 与 `activeRunnerPrCapability`，并附带 `commands` 和 `disclosure`，方便先看 scheduler auth 缺口、再看当前 runner/dev container 的 `gh auth status` 与 `gh pr create --dry-run` 能力；`--full` 或 `--raw` 才展开完整 `preflight`、工具、agent port、Git worktree、GitHub egress、repo/issue/PR 只读探测和观测原文。只报告 `GH_TOKEN`/`GITHUB_TOKEN` 是否存在和来源 key，不打印值。当 auth-broker 配置存在时，`tokenCoverage.source="auth-broker"`、`credentialSource="broker-issued-token"` 且 runner env token 不是成功前提；当仅 env token 存在时，`credentialSource="env-token"` 且 `authBroker.nextAction="use-env-token-until-auth-broker-live"`；两者都缺失时顶层 `ok=false`、`runnerDisposition=infra-blocked`、`degradedReason=auth-broker-needed`，`tokenCoverage.missing` 同时列出 `GH_TOKEN` 与 `GITHUB_TOKEN`，并输出 `authBroker.source="broker/auth-broker-needed"`、`capability.source="missing-token"`。该 `auth-missing` 的 scope 是 `scheduler-runner-env`，不能简化成“当前 active runner/dev container 不能创建 PR”；默认视图必须带 `scopeBoundary` 和 `activeRunnerPrCapability`。GitHub DNS/API 连接失败应归类为 `failureKind=github-transient`、`degradedReason=github-dns-api-transient`，并带 `retryable=true`、`commanderAction=retry-backoff-or-keep-running-if-heartbeat-fresh` 和有界 `githubTransient.failedProbes`；调用方应重试/退避，且在任务 heartbeat/trace 新鲜时继续监督，不把它当成 auth 缺失或 PR 语义失败。`prCapability` 是 runner-facing capability 摘要，必须包含目标分支、token/auth 来源、`systemGhBinaryRequiredForWrites=false`、UniDesk REST `bun scripts/cli.ts gh` 可用性、push dry-run/PR create dry-run 的 `writesRemote=false`、expected PR handoff、真实 PR 创建需要 commander 授权，以及 guarded `gh pr merge --dry-run` 预检路径；系统 `gh` binary 缺失只进入 `tools.systemGhBinary`，不得误判为 UniDesk REST `gh` CLI 不可用。`--remote` 在 runner-like 环境里不再依赖本地 `unidesk-backend-core`、`unidesk-database`、`baidu-netdisk-backend` 容器存在；这些缺失只作为本地观测证据。若远程控制面可达，则继续走远程控制面结果；若远程控制面不可达，则结构化返回 `failureKind=control-plane-missing` / `degradedReason=remote-control-plane-unreachable`，而不是把本地 `backend-core-container-missing` 当作最终阻塞。`--pr-create-dry-run` 不 POST GitHub，只证明 runner 内 PR body 生成、`scripts/cli.ts gh pr create --dry-run` 和 branch 参数形态可用；服务端创建权限仍以 token/auth broker、repo/issue/PR read、push dry-run 和最终授权后的真实 PR 创建结果为准。
 - `codex task <taskId>` 通过 Code Queue 私有代理按任务 ID 查询结构化审阅摘要；默认只返回任务身份、执行 Provider、工作目录、attempt 计数、原始 prompt、最终 response、最后错误和渐进披露命令，适合指挥官审阅完成未读任务且避免上下文爆炸。`--detail` 仍是有界详细摘要：默认只返回少量 attempt/tool 行、短 prompt/response/stderr/feedback 预览和 omitted/truncated 元数据；需要完整 prompt/response 文本或更多 tool/attempt 细节时再显式加 `--full`、`--tool-limit N`、`--trace` 或 `codex output`。该摘要读取默认由主 server `code-queue-mgr` 从 PostgreSQL 返回，不依赖 D601 `code-queue-read` Service 可用。
@@ -107,7 +107,7 @@ CI/CD、GitOps、rollout、artifact 发布、PR 合并后的 runtime lane 滚动
 - `codex dev-ready` 查询 Code Queue `/api/dev-ready` 并返回有界 readiness 摘要，包括工具、Docker、Codex config、SSH 和 `devReady.skills`。`devReady.skills` 只暴露 `UNIDESK_SKILLS_PATH`、是否存在、是否只读、skillCount、`cli-spec` 是否可见和修复建议，不输出宿主 auth/token 文件内容。
 - `codex judge <taskId> --attempt N [--dry-run] [--include-prompt]` 通过 Code Queue 私有代理按指定 attempt 单步复现 judge；这是执行面诊断入口，仍依赖 D601 scheduler/runner 侧的真实 judge builder、MiniMax 调用路径和执行环境。默认会真实调用 MiniMax，`--dry-run` 只返回 prompt/payload 大小、attempt 窗口和重建来源诊断，`--include-prompt` 仅用于本地深度排查。
 - `codex steer-confirm <taskId> --steer-id <id> [--raw]` 是只读 trace confirmation lookup。默认输出 `traceConfirmation.found/accepted/deliveryState/trace.seq/trace.at/promptChars/promptHash` 和 `delivery.status`，不回显 prompt；`--raw` 才附带原始 backend confirmation body。该命令用于处理 stable-proxy abort 后的 `deliveryUnconfirmed`，不要用重复 prompt 代替确认查询。
-- 旧 `codex steer` 已冻结；`codex steer-confirm` 只作为历史 trace confirmation lookup。新运行中纠偏使用 `bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin`，并用 `logs session/<sessionId>`、`events run/<runId>`、`result run/<runId> --command <commandId>` 和 `ack session/<sessionId>` 观察。
+- 旧 `codex steer` 已冻结；`codex steer-confirm` 只作为历史 trace confirmation lookup。新运行中纠偏使用 `bun scripts/cli.ts agentrun send session/<sessionId> --prompt-stdin`，并用 `logs session/<sessionId>`、`events run/<runId>`、`result run/<runId> --command <commandId>` 和 `ack session/<sessionId>` 观察。
 - `codex interrupt|cancel <taskId>` 通过 Code Queue 私有代理请求中断；running/judging 任务会请求 D601 当前 agent run 停止，queued/retry_wait 任务的取消也必须保持与 WebUI 相同代理路径，返回有界 task 摘要和后续查询命令。任何需要接触 active run 的动作仍属于 D601 执行面。
 - 旧 Code Queue 多队列 lane 现在是归档视图：`codex queues [--full|--all] [--limit N] [--page N|--offset N]` 只读展示历史 queue 摘要、activity、commanderConcurrency、counts 和 execution diagnostics。`queue create`、`queue merge`、`move` 等旧队列写入口冻结并返回 `legacy-code-queue-frozen`；AgentRun 新任务的排队、派发和取消必须使用 `agentrun create|apply|get|cancel`。
 - 所有旧 `codex` 历史查询、已读和残留 interrupt/cancel 命令必须走与 WebUI 相同的 backend-core 私有代理路径 `/api/microservices/code-queue/proxy/...`。旧 submit/steer/resume/queue mutation/move/workdir mutation 不得绕过冻结；若需要新任务或新 session 控制，使用 AgentRun 资源原语。
@@ -445,7 +445,7 @@ PATCH
 
 `--main-server-ip` 是一个全局前缀，必须放在需要透传的命令同一次调用中，例如 `bun scripts/cli.ts --main-server-ip 74.48.78.17 debug health`。默认传输是公网 frontend：本地 CLI 读取本仓库 `config.json` 中的 frontend 登录账号密码，登录 `http://<ip>:<frontendPort>/` 获取 HttpOnly session cookie，然后通过 frontend 的 `/api/*` 同源代理访问 backend-core 内网 API；因此计算节点只需要能访问公网 frontend，不需要主 server SSH key，也不需要打开 backend-core REST API 或 PostgreSQL 端口。
 
-默认 frontend 传输支持 `debug health`、`debug dispatch`、`debug task`、`artifact-registry status|health`、`ci publish-user-service --dry-run`、`microservice list/status/health/diagnostics/tunnel-self-test/proxy`、`decision upload/list/show/health`、`decision requirement list/upsert`、`decision diary import/list/history/months/show/edit/upsert`、`codex task <taskId>`、`codex tasks`、`codex unread`、`codex queues`、`codex output <taskId>`、`codex judge <taskId> --attempt N` 和 `ssh <PROVIDER_ID> <remote-command>`。`microservice status/health/diagnostics` 经 frontend 远程传输时也复用本地 CLI 的默认 compact summary，`microservice health code-queue` 只有显式 `--raw` 或 `--full` 才返回完整健康 body。运行中纠偏已切到 AgentRun `steer session/<sessionId>`；旧 `codex steer` 属于冻结写入口，不应通过 frontend 远程传输或旧 proxy 绕过。其中 `ssh` 的 remote frontend 传输使用 authenticated frontend `/ws/ssh` WebSocket 代理接入 backend-core SSH bridge，stdout/stderr 按字节流直通到调用端，不经过 `/api/dispatch`、`/api/tasks` 或 task JSON compact；frontend 运行时必须通过 `PROVIDER_TOKEN`/`UNIDESK_PROVIDER_TOKEN` 或 `PROVIDER_TOKEN_FILE`/`UNIDESK_PROVIDER_TOKEN_FILE` 读取 provider token，并且不能把 token 下发给 runner。因此 D601 Code Queue runner 内的 `tran G14 ...` 应与主 server 本机 `trans G14 ...` / `tran G14 ...` 在输出完整性上保持同一语义。非交互单进程命令优先 `trans D601 argv true`；`apply-patch`、stdin script、`py` 和旧 `apply-patch-v1` fallback 也走同一条 `/ws/ssh` 流式通道。交互式登录 shell 仍应在主 server 本机 CLI 使用，或显式切换到旧 SSH 传输后在主 server 上执行。当 backend-core、database、provider-dispatch 或 provider-host-ssh 缺失时，这些 read-only 预检必须返回结构化 `runnerDisposition=infra-blocked` 和缺失通道列表，而不是裸 `No such container`。若确实需要旧行为，可使用 `--main-server-key <key>` 或 `--main-server-transport ssh`，这时 CLI 会通过 SSH 登录主 server 的 `--main-server-root` 目录执行同一个 `bun scripts/cli.ts <command>`。
+默认 frontend 传输支持 `debug health`、`debug dispatch`、`debug task`、`artifact-registry status|health`、`ci publish-user-service --dry-run`、`microservice list/status/health/diagnostics/tunnel-self-test/proxy`、`decision upload/list/show/health`、`decision requirement list/upsert`、`decision diary import/list/history/months/show/edit/upsert`、`codex task <taskId>`、`codex tasks`、`codex unread`、`codex queues`、`codex output <taskId>`、`codex judge <taskId> --attempt N` 和 `ssh <PROVIDER_ID> <remote-command>`。`microservice status/health/diagnostics` 经 frontend 远程传输时也复用本地 CLI 的默认 compact summary，`microservice health code-queue` 只有显式 `--raw` 或 `--full` 才返回完整健康 body。运行中纠偏已切到 AgentRun `send session/<sessionId>`；旧 `codex steer` 属于冻结写入口，不应通过 frontend 远程传输或旧 proxy 绕过。其中 `ssh` 的 remote frontend 传输使用 authenticated frontend `/ws/ssh` WebSocket 代理接入 backend-core SSH bridge，stdout/stderr 按字节流直通到调用端，不经过 `/api/dispatch`、`/api/tasks` 或 task JSON compact；frontend 运行时必须通过 `PROVIDER_TOKEN`/`UNIDESK_PROVIDER_TOKEN` 或 `PROVIDER_TOKEN_FILE`/`UNIDESK_PROVIDER_TOKEN_FILE` 读取 provider token，并且不能把 token 下发给 runner。因此 D601 Code Queue runner 内的 `tran G14 ...` 应与主 server 本机 `trans G14 ...` / `tran G14 ...` 在输出完整性上保持同一语义。非交互单进程命令优先 `trans D601 argv true`；`apply-patch`、stdin script、`py` 和旧 `apply-patch-v1` fallback 也走同一条 `/ws/ssh` 流式通道。交互式登录 shell 仍应在主 server 本机 CLI 使用，或显式切换到旧 SSH 传输后在主 server 上执行。当 backend-core、database、provider-dispatch 或 provider-host-ssh 缺失时，这些 read-only 预检必须返回结构化 `runnerDisposition=infra-blocked` 和缺失通道列表，而不是裸 `No such container`。若确实需要旧行为，可使用 `--main-server-key <key>` 或 `--main-server-transport ssh`，这时 CLI 会通过 SSH 登录主 server 的 `--main-server-root` 目录执行同一个 `bun scripts/cli.ts <command>`。
 
 计算节点可以用该入口测试自身的远程升级闭环，而不需要在计算节点公开 core REST API 或 database。标准顺序是：先运行 `bun scripts/cli.ts --main-server-ip 74.48.78.17 debug health` 确认主 server 看到当前 Provider 在线，且该 Provider labels 中 `unideskCapabilities` 包含 `host.ssh`、`hostSshConfigured=true`、`hostSshKeyPresent=true`；再运行 `bun scripts/cli.ts --main-server-ip 74.48.78.17 debug dispatch <PROVIDER_ID> provider.upgrade --mode schedule --wait-ms 15000` 触发真实 `provider.upgrade`；随后再次运行 `debug health` 确认节点重新上线；最后运行 `bun scripts/cli.ts --main-server-ip 74.48.78.17 debug dispatch <PROVIDER_ID> host.ssh --wait-ms 15000` 和 `bun scripts/cli.ts --main-server-ip 74.48.78.17 ssh <PROVIDER_ID> hostname` 验证 SSH 透传能力。provider-gateway 新部署或升级后没有完成这组 remote CLI 自测，不能视为交付完成。
 

docs/reference/code-queue-supervision.md

@@ -93,7 +93,7 @@ HWLAB M3 口径使用同一分级：只读报告、fixture、LOCAL/DRY-RUN 和 d
 
 AgentRun 新派单和历史 Code Queue 审阅都按成本、可信度和 blast radius 分层：GPT-5.5/Codex 处理高风险和复杂任务，DeepSeek/OpenCode 处理中等复杂度且边界清晰的任务，MiniMax/OpenCode 处理简单、低权限、可复核任务，生产重启、密钥、数据库手工写入和运行中任务控制保留给指挥官或人工。
 
-当前新任务派发合同由 `bun scripts/cli.ts agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send` 资源原语暴露：`get tasks --queue commander` 查看指挥官队列，`create task --aipod Artificer --prompt-stdin` 或 `apply -f -` 创建任务，`dispatch task/<taskId>` 派发，`events/logs/result/ack/cancel/steer/send` 读取和控制 AgentRun task、run 与 session。UniDesk 是 render-only client：日常一次性 YAML/JSON 和 prompt 输入优先用 quoted heredoc/stdin，客户端按 `config/agentrun.yaml` 直连 AgentRun REST API 并保留 k8s 风格渲染；`--json-file`、`--prompt-file` 和 `--runner-json-file` 只是客户端输入来源。该路径不经过 HWLAB runtime、SSH official CLI 或旧 bridge wrapper，不做旧 Code Queue 双写，也不迁移旧历史。
+当前新任务派发合同由 `bun scripts/cli.ts agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send` 资源原语暴露：`get tasks --queue commander` 查看指挥官队列，`create task --aipod Artificer --prompt-stdin` 或 `apply -f -` 创建任务，`dispatch task/<taskId>` 派发，`events/logs/result/ack/cancel/send` 读取和控制 AgentRun task、run 与 session。UniDesk 是 render-only client：日常一次性 YAML/JSON 和 prompt 输入优先用 quoted heredoc/stdin，客户端按 `config/agentrun.yaml` 直连 AgentRun REST API 并保留 k8s 风格渲染；`send session/<sessionId>` 是唯一用户级 session follow-up 写入口，服务端按 durable session/run/command 状态自动决定内部 `steer` 或新 `turn`，旧 CLI `turn/steer` 路径不保留兼容。`--json-file`、`--prompt-file` 和 `--runner-json-file` 只是客户端输入来源。该路径不经过 HWLAB runtime、SSH official CLI 或旧 bridge wrapper，不做旧 Code Queue 双写，也不迁移旧历史。
 旧 `codex submit/enqueue`、`codex steer`、`codex resume`、旧 queue mutation、task move 和旧 workdir mutation 已冻结。CLI 必须返回 `ok=false`、`frozen=true`、`degradedReason=legacy-code-queue-frozen` 和 AgentRun 替代命令；服务端旧 API 写入口必须返回 410。旧 `codex task/tasks/output/read/unread/queues` 继续作为历史归档和只读排障入口，`codex interrupt|cancel` 只用于停止残留旧任务。
 
 新任务模型由 AgentRun task payload 和 AgentRun runtime 配置决定；旧 Code Queue 的 `CODE_QUEUE_MODELS` 只作为历史任务审阅和残留运行面配置参考，长期合同至少包含 GPT-5.5、GPT-5.4、GPT-5.4 Mini、DeepSeek Chat、MiniMax M3 和 MiniMax M2.7 两路并行配置；`deepseek`/`deepseek-chat`、`minimax-m3` 与 `minimax-m2.7` 会走 OpenCode port，其余模型走 Codex port。PROD 集群把 `MINIMAX_MODEL` 切到 `MiniMax-M3`（M3 是新任务的默认 provider model），judge 与 opencode 跟随；M2.7 仍然作为并行配置存在，切换只需把 `MINIMAX_MODEL` 改成 `MiniMax-M2.7` 后 rollout restart。两者不存在自动 fallback 关系：M3 任务失败不会自动改派 M2.7，task 要用 M2.7 必须显式 `--model minimax-m2.7`。只有当执行面 `/health` 或等价配置已经显示 DeepSeek 模型可用、并完成轻量 runner smoke 后，才允许真实提交 `--model deepseek-chat`。
@@ -268,15 +268,15 @@ bun scripts/cli.ts codex pr-preflight --remote --issue <issue-number>
 
 ### Runner Resume 收口
 
-PR 小修、冲突、rebase、补测和 reviewer feedback 的新执行入口是 AgentRun 资源原语。仍在 AgentRun session 内的工作优先使用 `bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin` 或 `send session/<sessionId>`；已沉淀成新工作项时使用 `bun scripts/cli.ts agentrun create task --aipod Artificer --prompt-stdin` 或 `agentrun apply -f -`。旧 `codex resume` 已冻结，不再作为 follow-up turn 入口。
+PR 小修、冲突、rebase、补测和 reviewer feedback 的新执行入口是 AgentRun 资源原语。仍在 AgentRun session 内的工作只使用 `bun scripts/cli.ts agentrun send session/<sessionId> --prompt-stdin`；已沉淀成新工作项时使用 `bun scripts/cli.ts agentrun create task --aipod Artificer --prompt-stdin` 或 `agentrun apply -f -`。旧 `codex resume` 已冻结，不再作为 follow-up 入口。
 
 旧 Code Queue task 只保留历史审阅和残留停止；需要基于旧任务产出继续推进时，在 AgentRun payload 中显式引用旧 task id、PR/branch 和审阅结论，而不是把旧 task 重新入队、resume 或 double-write。
 
 AgentRun turn 的 timeout 监督按无响应空闲时间处理，而不是固定 wall-clock `backend-timeout`。只要 result/session liveness 的 `lastActivityAt`、`lastActivitySeq` 或 events 仍在刷新，指挥官应继续轮询，不应 interrupt、重派或判定 backend 死亡。若 command 因 idle timeout、provider stream disconnect、runner stdio inactive 或其他非业务终态停止，指挥官本人必须先读 `result`、`events` 或 `logs/trace`，确认最后有效工具输出、已完成修改、失败原因和下一步，再把这个摘要作为后续 prompt 发送给 Artificer；不能要求 Artificer 自己去猜旧 trace，也不能因为一次 timeout 就停止整个交付。
 
-Artificer 默认应携带可续跑 `sessionRef`。仍有 `sessionId` 的 follow-up、补测、reviewer feedback 和 timeout 恢复，优先使用 `bun scripts/cli.ts agentrun send session/<sessionId> --aipod Artificer --prompt-stdin` 或 `steer session/<sessionId>` 继续同一个 session；只有历史任务没有 `sessionRef`、session 已 evicted、或同 session 已证明不可恢复时，才创建新 AgentRun task。新 task 必须在 prompt 中写明“这是基于旧 task/trace 的 manager-read continuation”，并记录旧 task/run/branch、管理者读 trace 得出的当前状态和下一步。
+Artificer 默认应携带可续跑 `sessionRef`。仍有 `sessionId` 的 follow-up、补测、reviewer feedback 和 timeout 恢复，只使用 `bun scripts/cli.ts agentrun send session/<sessionId> --aipod Artificer --prompt-stdin` 继续同一个 session；只有历史任务没有 `sessionRef`、session 已 evicted、或同 session 已证明不可恢复时，才创建新 AgentRun task。新 task 必须在 prompt 中写明“这是基于旧 task/trace 的 manager-read continuation”，并记录旧 task/run/branch、管理者读 trace 得出的当前状态和下一步。
 
-replacement runner 只用于方向明显错误、质量不可接受、原 task 上下文不可恢复、原分支/PR 已废弃，或 AgentRun reuse/steer 已证明无法继续的情况。关闭或替换旧 PR 时必须在 PR/body/final response 中说明 superseded/replacement 关系，避免 competing branch 扩散。
+replacement runner 只用于方向明显错误、质量不可接受、原 task 上下文不可恢复、原分支/PR 已废弃，或 AgentRun 同 session `send` 已证明无法继续的情况。关闭或替换旧 PR 时必须在 PR/body/final response 中说明 superseded/replacement 关系，避免 competing branch 扩散。
 
 ## 监控
 
@@ -294,7 +294,7 @@ replacement runner 只用于方向明显错误、质量不可接受、原 task
 - `bun scripts/cli.ts codex tasks --status succeeded --unread --limit N`：按具体终态过滤监督结果；不支持的 status filter 必须显式失败，不能扩大为未过滤结果。
 - `bun scripts/cli.ts codex task <taskId>`：默认只查看原始 prompt、最终 response、最后错误和 drill-down 命令，这是完成未读任务审阅的第一步。
 - 当默认审阅摘要不足时，再逐级使用 `bun scripts/cli.ts codex task <taskId> --detail`、`bun scripts/cli.ts codex task <taskId> --trace --limit N` 或 `codex output`。
-- `bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin`：对 AgentRun 中仍可继续的 session 追加修正；旧 `codex resume` 已冻结。
+- `bun scripts/cli.ts agentrun send session/<sessionId> --prompt-stdin`：对 AgentRun 中仍可继续的 session 追加修正；服务端按运行状态决定内部 `steer` 或新 `turn`，旧 `codex resume` 已冻结。
 - 当 master 控制面状态和 D601 scheduler 状态看起来分裂时，使用 `docs/reference/observability.md` 中的活性规则判断。
 
 默认 commander/supervisor 视图必须保持低噪声。commander 视图用于回答“现在需要处理什么”，supervisor 视图用于看分区小页和红线细节。commander 的 `activeRunners.count` 是指挥官 active runner 计数，supervisor 的 `activeRunning.count` 是 running+judging 状态计数；两者都必须标明 exact/source，不能把返回行数当成并发总数。`activeRunning.count` 来源是 queue summary 的 status counts 时 `activeRunning.exact=true`，用于 redline 判断；`activeRunning.rowPage.returned` / `running.returned` 只表示本次返回的紧凑任务行。`activeRunning.redline` 必须写明 `countField`、routine target、burst redline、hard redline、`state` 和 `decisionReady`；只有 `decisionReady=true` 时，才能直接用该 count 做红线/补派判断。commander 的 `attention.items` 只返回最需要处理的有界任务，`attention.total/returned/omitted` 必须保留省略计数；`sections.recentCompleted` 不得重复 `sections.terminalUnread` 的未读终态。`running`、`completedUnread` 和 `queued` 即使传入较大的 `--limit`，默认也只返回一个很小的有界页，并通过 section `commands.next` 继续分页；`--limit` 保留为扫描/分页预算和 full view 返回预算，不得让一次 commander/supervisor 调用输出几十条肥行。每个任务行只应带 task id 和必要摘要，`show`、`detail`、`trace`、`output`、`full`、`read` 使用 section template 或 row commands 表达，让下一步渐进披露动作明确且不重复；默认不得嵌入完整 queue 列表、完整 final response、raw output 页或完整 trace 行。`recentCompleted` 必须默认限量，且不得重复 `completedUnread` 里的未读终态，避免完成历史把当前 running、阻塞和未读审阅挤出视野；需要完整当前页时显式使用 `--view full`。`executionDiagnostics` 只能展示有界 task-id/reason 预览、总数、截断标记和 omitted counts；需要全量诊断时使用输出中的 raw command。`commands.read` 只是在人工审阅后的建议命令，listing 命令绝不能自动执行。
@@ -373,8 +373,8 @@ D601 artifact registry 的 systemd unit inactive 不等于 D601 全局离线。
 只有存在明确理由时才干预。
 
 - 如果任务还在运行且 trace 或 scheduler heartbeat 新鲜，应引导而不是 interrupt。
-- 对 AgentRun 运行中 session 的引导应优先使用正式 CLI：`bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin`，再用 `logs/events/result/ack` 确认。旧 `codex steer` 已冻结，只保留历史 trace confirmation 查询。
-- 真实 steer 输出必须保持低噪声：成功显示 `steer.status`、`steer.deliveryState`、`steer.steerId`、有界 `traceConfirmation` 和后续命令，不回显 prompt 或完整 task state；失败默认不带 request body、不带 upstream body preview，也不带 raw response，需要上游预览或原始失败对象时显式重跑 `--full` 或 `--raw`。`deliveryState=accepted` 表示 backend 已接受；`not_accepted` 表示任务状态/权限/输入未接受；`accepted_response_timeout` 表示 stable proxy 响应超时但 trace confirmation 找到该 `steerId`；`unknown` 表示响应路径失败且确认查询仍未证明接受。
+- 对 AgentRun 运行中 session 的引导只使用正式 CLI：`bun scripts/cli.ts agentrun send session/<sessionId> --prompt-stdin`，再用 `logs/events/result/ack` 确认。旧 `codex steer` 已冻结，只保留历史 trace confirmation 查询。
+- 真实 `send` 输出必须保持低噪声并直接显示 `dryRun`、`mutation`、`decision`、`internalCommandType`、run/command/runnerjob 摘要和后续命令，不回显 prompt 或完整 task state；失败默认不带 request body、不带 upstream body preview，也不带 raw response，需要上游预览或原始失败对象时显式重跑 `--full` 或 `--raw`。
 - 旧 Code Queue 的 provider tunnel 失败只作为历史运行面诊断线索；新任务控制面失败优先按 AgentRun `describe task`、`events`、`logs`、`result`、G14 `agentrun-v01` manager 和 runner job 证据分流。
 - 新 AgentRun 任务失败分流以 AgentRun queue/session/runner-job 返回字段为准。旧 Code Queue `.data.diagnostics.reason` 只用于历史任务和残留运行面，不再引导新 `codex submit/steer/resume`。
 - 如果任务进入终态但缺少必要验收证据，应使用聚焦 continuation prompt retry 同一任务。

scripts/src/agentrun.ts

@@ -39,7 +39,7 @@ const gitMirrorRepositories: readonly GitMirrorRepositorySpec[] = [
 
 export function agentRunHelp(): unknown {
   return {
-    command: "agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send|explain",
+    command: "agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send|explain",
     output: "human by default; use -o json|yaml or --raw for machine/debug output",
     usage: [
       "bun scripts/cli.ts agentrun get tasks --queue commander --limit 20",
@@ -54,7 +54,6 @@ export function agentRunHelp(): unknown {
       "bun scripts/cli.ts agentrun dispatch task/<taskId>",
       "bun scripts/cli.ts agentrun create task --aipod Artificer --prompt-stdin --idempotency-key <key>",
       "bun scripts/cli.ts agentrun apply -f - --dry-run",
-      "bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin",
       "bun scripts/cli.ts agentrun send session/<sessionId> --aipod Artificer --prompt-stdin",
       "bun scripts/cli.ts agentrun explain task",
       "bun scripts/cli.ts agentrun control-plane status",
@@ -78,7 +77,7 @@ export function agentRunHelp(): unknown {
     ],
     resources: ["task/qt", "run", "command/cmd", "runnerjob/rjob", "session/ses", "aipodspec/aps"],
     description: "Operate AgentRun v0.1 through Kubernetes-style resource verbs. Human output is compact by default; -o json|yaml returns the UniDesk render-only client schema, and --raw exposes the direct AgentRun REST envelope.",
-    legacyCompatibility: "queue/runs/commands/runner/sessions/aipod-specs remain as compatibility groups backed by direct HTTP; new commander work should use get/describe/events/logs/result/ack/cancel/dispatch/create/apply/steer/send.",
+    legacyCompatibility: "queue/runs/commands/runner/sessions/aipod-specs remain as compatibility groups backed by direct HTTP; new commander work should use get/describe/events/logs/result/ack/cancel/dispatch/create/apply/send. sessions turn/steer are removed; use send only.",
   };
 }
 
@@ -150,7 +149,6 @@ function isResourceVerb(value: string | undefined): value is AgentRunResourceVer
     || value === "dispatch"
     || value === "create"
     || value === "apply"
-    || value === "steer"
     || value === "send"
     || value === "explain";
 }
@@ -211,11 +209,8 @@ function agentRunHelpText(args: string[]): string {
   if (verb === "apply") {
     return "Usage: bun scripts/cli.ts agentrun apply -f task.yaml|json|- [--dry-run]\nTask manifests use kind: Task and spec: <AgentRun task payload>.";
   }
-  if (verb === "steer") {
-    return "Usage: bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin";
-  }
   if (verb === "send") {
-    return "Usage: bun scripts/cli.ts agentrun send session/<sessionId> --aipod Artificer --prompt-stdin";
+    return "Usage: bun scripts/cli.ts agentrun send session/<sessionId> --aipod Artificer --prompt-stdin\nThe server decides whether this becomes an internal steer or a new turn from session state.";
   }
   if (verb === "explain") return agentRunExplain(kind ?? "task");
   if (verb === "control-plane") {
@@ -254,7 +249,7 @@ function agentRunHelpText(args: string[]): string {
   return [
     "Usage: bun scripts/cli.ts agentrun <verb> <resource> [options]",
     "",
-    "Verbs: get, describe, events, logs, result, ack, cancel, dispatch, create, apply, steer, send, explain",
+    "Verbs: get, describe, events, logs, result, ack, cancel, dispatch, create, apply, send, explain",
     "Resources: task/qt, run, command/cmd, runnerjob/rjob, session/ses, aipodspec/aps",
     "",
     "Common:",
@@ -299,8 +294,7 @@ async function runAgentRunResourceCommand(config: UniDeskConfig | null, verb: Ag
     if (verb === "dispatch") return await resourceDispatch(config, command, action, bridgeActionArgs, options);
     if (verb === "create") return await resourceCreate(config, command, action, bridgeActionArgs, options);
     if (verb === "apply") return await resourceApply(config, command, bridgeActionArgs, options);
-    if (verb === "steer") return await resourceSessionPromptCommand(config, command, "steer", action, bridgeActionArgs, options);
-    if (verb === "send") return await resourceSessionPromptCommand(config, command, "turn", action, bridgeActionArgs, options);
+    if (verb === "send") return await resourceSessionPromptCommand(config, command, action, bridgeActionArgs, options);
   } catch (error) {
     if (error instanceof AgentRunRestError) return renderAgentRunRestError(command, error, options);
     return renderedCliResult(false, command, `Error: ${error instanceof Error ? error.message : String(error)}`);
@@ -620,12 +614,12 @@ async function resourceApply(config: UniDeskConfig | null, command: string, args
   return renderMutationSummary(command, result, options, `${options.dryRun ? "Dry-run applied" : "Applied"} task manifest`, options.dryRun ? [rerunWithoutDryRun(command)] : undefined);
 }
 
-async function resourceSessionPromptCommand(config: UniDeskConfig | null, command: string, officialAction: "steer" | "turn", action: string | undefined, args: string[], options: AgentRunResourceOptions): Promise<RenderedCliResult> {
+async function resourceSessionPromptCommand(config: UniDeskConfig | null, command: string, action: string | undefined, args: string[], options: AgentRunResourceOptions): Promise<RenderedCliResult> {
   const ref = parseResourceRef(action, args, "session");
-  if (ref.kind !== "session") throw new Error(`${officialAction === "steer" ? "steer" : "send"} requires session/<sessionId>`);
-  const sessionArgs = [officialAction, ref.name, ...stripLeadingResource(args, ref.name)];
+  if (ref.kind !== "session") throw new Error("send requires session/<sessionId>");
+  const sessionArgs = ["send", ref.name, ...stripLeadingResource(args, ref.name)];
   const result = await runAgentRunRestCommand(config, "sessions", sessionArgs);
-  return renderMutationSummary(command, result, options, officialAction === "steer" ? "Steer submitted" : "Session turn submitted");
+  return renderMutationSummary(command, result, options, options.dryRun ? "Session send plan" : "Session send submitted");
 }
 
 function renderedCliResult(ok: boolean, command: string, renderedText: string, contentType: RenderedCliResult["contentType"] = "text/plain"): RenderedCliResult {
@@ -675,7 +669,20 @@ function renderResultSummary(command: string, raw: Record<string, unknown>, opti
     `State:  ${displayValue(data.state ?? data.status ?? data.terminalStatus ?? "-")}`,
     `OK:     ${String(raw.ok !== false)}`,
   ];
-  const final = stringOrNull(data.finalResponse) ?? stringOrNull(data.output) ?? stringOrNull(data.summary) ?? stringOrNull(data.result);
+  const authority = stringOrNull(data.finalResponseAuthority);
+  const fallback = typeof data.finalResponseFallback === "boolean" ? data.finalResponseFallback : null;
+  const needsContinuation = typeof data.needsContinuation === "boolean" ? data.needsContinuation : null;
+  if (authority !== null) lines.push(`Authority: ${authority}`);
+  if (fallback !== null) lines.push(`Fallback: ${String(fallback)}`);
+  if (needsContinuation !== null) lines.push(`NeedsContinuation: ${String(needsContinuation)}`);
+  const finalRecord = record(data.finalResponse);
+  const final = stringOrNull(data.finalResponse)
+    ?? stringOrNull(finalRecord.content)
+    ?? stringOrNull(finalRecord.text)
+    ?? stringOrNull(finalRecord.message)
+    ?? stringOrNull(data.output)
+    ?? stringOrNull(data.summary)
+    ?? stringOrNull(data.result);
   if (final !== null) lines.push("", truncateMultiline(final, options.fullText ? 8000 : 1600));
   else {
     const failure = renderFailureLines(data);
@@ -694,6 +701,12 @@ function renderMutationSummary(command: string, raw: Record<string, unknown>, op
     `OK: ${String(raw.ok !== false)}`,
   ];
   if (id !== null) lines.push(`Name: ${id}`);
+  const decision = stringOrNull(data.decision);
+  const internalCommandType = stringOrNull(data.internalCommandType);
+  if (data.dryRun !== undefined) lines.push(`DryRun: ${String(data.dryRun)}`);
+  if (data.mutation !== undefined) lines.push(`Mutation: ${String(data.mutation)}`);
+  if (decision !== null) lines.push(`Decision: ${decision}`);
+  if (internalCommandType !== null) lines.push(`InternalCommandType: ${internalCommandType}`);
   const next = record(raw.next ?? data.next);
   const nextLines = (overrideNextLines ?? Object.values(next).map(String)).filter((line) => line.length > 0).slice(0, 5);
   if (nextLines.length > 0) lines.push("", "Next:", ...nextLines.map((line) => `  ${line}`));
@@ -3342,8 +3355,7 @@ async function runAgentRunSessionsRest(action: string | undefined, id: string |
     if (agentRunHasFlag(args, "dry-run")) return agentRunDryRunPlan("session-cancel", `/api/v1/sessions/${encodeURIComponent(id)}/control`, body, `bun scripts/cli.ts agentrun sessions cancel ${id}`);
     return await agentRunRestRequest("agentrun sessions cancel", "POST", `/api/v1/sessions/${encodeURIComponent(id)}/control`, body);
   }
-  if (action === "steer" && id) return await sessionSteerRest(id, args);
-  if (action === "turn") return await sessionTurnRest(id ?? null, args);
+  if (action === "send" && id) return await sessionSendRest(id, args);
   throw new AgentRunRestError("validation-failed", `unsupported sessions command: ${[action, id].filter(Boolean).join(" ") || "(empty)"}`);
 }
 
@@ -3436,64 +3448,85 @@ async function mutateQueueTaskRest(action: string, taskId: string, suffix: strin
   return await agentRunRestRequest(`agentrun queue ${suffix}`, "POST", pathValue, body);
 }
 
-async function sessionSteerRest(sessionId: string, args: string[]): Promise<Record<string, unknown>> {
-  const session = record(innerData(await agentRunRestRequest("agentrun sessions show", "GET", `/api/v1/sessions/${encodeURIComponent(sessionId)}${agentRunQuery(args, ["reader-id"])}`)));
-  const runId = stringOrNull(session.activeRunId) ?? stringOrNull(session.lastRunId);
-  if (runId === null) throw new AgentRunRestError("validation-failed", `session ${sessionId} has no run to steer`);
-  const prompt = readPromptFromArgs(args, 2);
-  const body: Record<string, unknown> = { type: "steer", payload: { prompt } };
-  const idempotencyKey = agentRunOption(args, "idempotency-key");
-  if (idempotencyKey) body.idempotencyKey = idempotencyKey;
-  const command = await agentRunRestRequest("agentrun sessions steer", "POST", `/api/v1/runs/${encodeURIComponent(runId)}/commands`, body);
-  return { ok: command.ok !== false, command: "agentrun sessions steer", data: { action: "session-steer", sessionId, runId, command: innerData(command) }, bridge: command.bridge };
+async function sessionSendRest(sessionId: string, args: string[]): Promise<Record<string, unknown>> {
+  const aipod = agentRunOption(args, "aipod") ?? agentRunOption(args, "aipod-spec");
+  if (aipod) return await sessionSendWithAipodRest(sessionId, aipod, args);
+  const input = await optionalJsonBody(args);
+  const prompt = optionalPromptFromArgs(args, 2);
+  const sendBody = await sessionSendBodyFromRunInput(sessionId, args, input, sessionSendPayloadFromInput(input, prompt));
+  const response = await agentRunRestRequest("agentrun sessions send", "POST", `/api/v1/sessions/${encodeURIComponent(sessionId)}/send`, sendBody);
+  return { ok: response.ok !== false, command: "agentrun sessions send", data: innerData(response), bridge: response.bridge };
 }
 
-async function sessionTurnRest(positionalSessionId: string | null, args: string[]): Promise<Record<string, unknown>> {
-  const aipod = agentRunOption(args, "aipod") ?? agentRunOption(args, "aipod-spec");
-  if (aipod) return await sessionTurnWithAipodRest(positionalSessionId, aipod, args);
-  const sessionId = positionalSessionId ?? agentRunOption(args, "session-id") ?? newAgentRunSessionId();
-  const body = await optionalJsonBody(args);
-  const profile = agentRunOption(args, "profile") ?? agentRunOption(args, "backend-profile") ?? stringOrNull(body.backendProfile) ?? "codex";
-  const prompt = readPromptFromArgs(args, positionalSessionId ? 2 : 1);
-  body.tenantId = agentRunOption(args, "tenant-id") ?? stringOrNull(body.tenantId) ?? "unidesk";
-  body.projectId = agentRunOption(args, "project-id") ?? stringOrNull(body.projectId) ?? "default";
+function sessionSendPayloadFromInput(input: Record<string, unknown>, prompt: string | null): Record<string, unknown> {
+  const payload = record(input.payload);
+  if (Object.keys(payload).length > 0) return payload;
+  const text = prompt ?? stringOrNull(input.prompt) ?? stringOrNull(input.message) ?? stringOrNull(input.text);
+  if (text === null) throw new AgentRunRestError("validation-failed", "send requires payload or non-empty prompt/message/text; use --prompt-stdin, --prompt-file, --prompt, a trailing prompt, or JSON payload");
+  return { prompt: text };
+}
+
+async function sessionSendBodyFromRunInput(sessionId: string, args: string[], input: Record<string, unknown>, payload: Record<string, unknown>): Promise<Record<string, unknown>> {
+  const fullSendBody = isExplicitSessionSendBody(input);
+  const runInput = fullSendBody ? record(input.run ?? input.runBase) : input;
+  const runBody = await sessionRunBodyFromArgs(sessionId, args, runInput);
+  const runnerDefaults = fullSendBody ? record(input.runnerJob) : {};
+  const runnerJob = await sessionRunnerJobBody(args, runnerDefaults);
+  const body: Record<string, unknown> = fullSendBody ? { ...input } : {};
+  body.run = runBody;
+  body.payload = record(input.payload);
+  if (Object.keys(record(body.payload)).length === 0) body.payload = payload;
+  body.createRunnerJob = agentRunHasFlag(args, "no-runner-job") ? false : input.createRunnerJob !== false;
+  body.runnerJob = runnerJob;
+  body.dryRun = agentRunHasFlag(args, "dry-run");
+  const commandIdempotencyKey = agentRunOption(args, "command-idempotency-key") ?? agentRunOption(args, "idempotency-key");
+  if (commandIdempotencyKey) body.commandIdempotencyKey = commandIdempotencyKey;
+  return body;
+}
+
+function isExplicitSessionSendBody(input: Record<string, unknown>): boolean {
+  return isRecord(input.run)
+    || isRecord(input.runBase)
+    || isRecord(input.payload)
+    || isRecord(input.runnerJob)
+    || input.createRunnerJob !== undefined
+    || input.commandIdempotencyKey !== undefined;
+}
+
+async function sessionRunBodyFromArgs(sessionId: string, args: string[], input: Record<string, unknown>): Promise<Record<string, unknown>> {
+  const existing = await fetchAgentRunSessionOrNull(sessionId, args);
+  const profile = agentRunOption(args, "profile")
+    ?? agentRunOption(args, "backend-profile")
+    ?? stringOrNull(input.backendProfile)
+    ?? stringOrNull(existing?.backendProfile)
+    ?? "codex";
+  const body: Record<string, unknown> = { ...input };
+  body.tenantId = agentRunOption(args, "tenant-id") ?? stringOrNull(body.tenantId) ?? stringOrNull(existing?.tenantId) ?? "unidesk";
+  body.projectId = agentRunOption(args, "project-id") ?? stringOrNull(body.projectId) ?? stringOrNull(existing?.projectId) ?? "default";
   body.providerId = agentRunOption(args, "provider-id") ?? stringOrNull(body.providerId) ?? "G14";
   body.backendProfile = profile;
   body.workspaceRef = jsonObjectOption(args, "workspace-json") ?? record(body.workspaceRef);
   if (Object.keys(record(body.workspaceRef)).length === 0) body.workspaceRef = { kind: "opaque", path: "." };
   body.executionPolicy = jsonObjectOption(args, "execution-policy-json") ?? record(body.executionPolicy);
   if (Object.keys(record(body.executionPolicy)).length === 0) body.executionPolicy = defaultAgentRunExecutionPolicy(profile);
-  const sessionRef = record(body.sessionRef);
-  const metadata = record(sessionRef.metadata);
+  const inheritedSessionRef = existing === null ? {} : {
+    conversationId: existing.conversationId,
+    threadId: existing.threadId,
+    metadata: existing.metadata,
+  };
+  const sessionRef = { ...inheritedSessionRef, ...record(body.sessionRef) };
+  const metadata = { ...record(inheritedSessionRef.metadata), ...record(sessionRef.metadata) };
   const title = agentRunOption(args, "title");
   if (title) metadata.title = title;
   body.sessionRef = { ...sessionRef, sessionId, metadata };
-  await ensureAgentRunSession(sessionId, body);
-  const run = await agentRunRestRequest("agentrun runs create", "POST", "/api/v1/runs", body);
-  const runData = record(innerData(run));
-  const runId = stringOrNull(runData.id) ?? requiredContext("session turn", "run id returned by AgentRun server");
-  const commandBody: Record<string, unknown> = { type: "turn", payload: { prompt } };
-  const commandIdempotencyKey = agentRunOption(args, "command-idempotency-key") ?? agentRunOption(args, "idempotency-key");
-  if (commandIdempotencyKey) commandBody.idempotencyKey = commandIdempotencyKey;
-  const command = await agentRunRestRequest("agentrun commands create", "POST", `/api/v1/runs/${encodeURIComponent(runId)}/commands`, commandBody);
-  const commandId = stringOrNull(record(innerData(command)).id);
-  let runnerJob: Record<string, unknown> | null = null;
-  if (!agentRunHasFlag(args, "no-runner-job") && commandId !== null) {
-    const runnerBody = await optionalRunnerJsonBody(args);
-    runnerBody.commandId = commandId;
-    copyAgentRunOptions(args, runnerBody, ["image", "namespace", "attempt-id", "runner-id", "source-commit", "service-account-name"]);
-    runnerJob = await agentRunRestRequest("agentrun runner job", "POST", `/api/v1/runs/${encodeURIComponent(runId)}/runner-jobs`, runnerBody);
-  }
-  return { ok: true, command: "agentrun sessions turn", data: { action: "session-turn", sessionId, profile, run: innerData(run), command: innerData(command), runnerJob: runnerJob ? innerData(runnerJob) : null, valuesPrinted: false }, bridge: runnerJob?.bridge ?? command.bridge ?? run.bridge };
+  return body;
 }
 
-async function sessionTurnWithAipodRest(positionalSessionId: string | null, aipod: string, args: string[]): Promise<Record<string, unknown>> {
-  const sessionId = positionalSessionId ?? agentRunOption(args, "session-id") ?? newAgentRunSessionId();
-  const rendered = await agentRunRestRequest("agentrun aipod-specs render", "POST", `/api/v1/aipod-specs/${encodeURIComponent(aipod)}/render`, await aipodRenderInputFromArgs(args, positionalSessionId ? 2 : 1, { sessionId }));
+async function sessionSendWithAipodRest(sessionId: string, aipod: string, args: string[]): Promise<Record<string, unknown>> {
+  const rendered = await agentRunRestRequest("agentrun aipod-specs render", "POST", `/api/v1/aipod-specs/${encodeURIComponent(aipod)}/render`, await aipodRenderInputFromArgs(args, 2, { sessionId }));
   const renderedData = record(innerData(rendered));
   const task = record(renderedData.queueTask);
   if (Object.keys(task).length === 0) throw new AgentRunRestError("schema-mismatch", `aipod-spec ${aipod} render did not return queueTask`);
-  await ensureAgentRunSession(sessionId, task);
   const sessionRef = record(task.sessionRef);
   const metadata = record(sessionRef.metadata);
   const title = agentRunOption(args, "title") ?? stringOrNull(task.title);
@@ -3509,37 +3542,38 @@ async function sessionTurnWithAipodRest(positionalSessionId: string | null, aipo
     resourceBundleRef: task.resourceBundleRef,
     traceSink: { kind: "aipod-session", aipod, sessionId, valuesPrinted: false },
   };
-  const run = await agentRunRestRequest("agentrun runs create", "POST", "/api/v1/runs", runBody);
-  const runId = stringOrNull(record(innerData(run)).id) ?? requiredContext("session turn", "run id returned by AgentRun server");
-  const commandBody: Record<string, unknown> = { type: "turn", payload: task.payload };
+  const runnerDefaults = record(record(renderedData.dispatchDefaults).runnerJob);
+  const sendBody: Record<string, unknown> = {
+    run: runBody,
+    payload: task.payload,
+    createRunnerJob: !agentRunHasFlag(args, "no-runner-job"),
+    runnerJob: await sessionRunnerJobBody(args, runnerDefaults),
+    dryRun: agentRunHasFlag(args, "dry-run"),
+  };
   const commandIdempotencyKey = agentRunOption(args, "command-idempotency-key") ?? agentRunOption(args, "idempotency-key");
-  if (commandIdempotencyKey) commandBody.idempotencyKey = commandIdempotencyKey;
-  const command = await agentRunRestRequest("agentrun commands create", "POST", `/api/v1/runs/${encodeURIComponent(runId)}/commands`, commandBody);
-  const commandId = stringOrNull(record(innerData(command)).id);
-  let runnerJob: Record<string, unknown> | null = null;
-  if (!agentRunHasFlag(args, "no-runner-job") && commandId !== null) {
-    const runnerDefaults = record(record(renderedData.dispatchDefaults).runnerJob);
-    const runnerBody = { ...runnerDefaults, ...(await optionalRunnerJsonBody(args)), commandId };
-    copyAgentRunOptions(args, runnerBody, ["image", "namespace", "attempt-id", "runner-id", "source-commit", "service-account-name"]);
-    runnerJob = await agentRunRestRequest("agentrun runner job", "POST", `/api/v1/runs/${encodeURIComponent(runId)}/runner-jobs`, runnerBody);
-  }
-  return { ok: true, command: "agentrun sessions turn", data: { action: "session-turn", aipod, sessionId, profile: String(task.backendProfile ?? ""), run: innerData(run), command: innerData(command), runnerJob: runnerJob ? innerData(runnerJob) : null, valuesPrinted: false }, bridge: runnerJob?.bridge ?? command.bridge ?? run.bridge };
+  if (commandIdempotencyKey) sendBody.commandIdempotencyKey = commandIdempotencyKey;
+  const response = await agentRunRestRequest("agentrun sessions send", "POST", `/api/v1/sessions/${encodeURIComponent(sessionId)}/send`, sendBody);
+  const data = record(innerData(response));
+  return { ok: response.ok !== false, command: "agentrun sessions send", data: { ...data, aipod, profile: String(task.backendProfile ?? ""), valuesPrinted: false }, bridge: response.bridge };
 }
 
-async function ensureAgentRunSession(sessionId: string, source: Record<string, unknown>): Promise<void> {
+async function fetchAgentRunSessionOrNull(sessionId: string, args: string[]): Promise<Record<string, unknown> | null> {
   try {
-    await agentRunRestRequest("agentrun sessions storage", "GET", `/api/v1/sessions/${encodeURIComponent(sessionId)}/storage`);
-    return;
-  } catch {
-    // Missing storage is recovered by creating the session; other failures will surface on create.
+    return record(innerData(await agentRunRestRequest("agentrun sessions show", "GET", `/api/v1/sessions/${encodeURIComponent(sessionId)}${agentRunQuery(args, ["reader-id"])}`)));
+  } catch (error) {
+    if (error instanceof AgentRunRestError && error.httpStatus === 404) return null;
+    throw error;
   }
-  await agentRunRestRequest("agentrun sessions create", "POST", "/api/v1/sessions", {
-    sessionId,
-    tenantId: source.tenantId ?? "unidesk",
-    projectId: source.projectId ?? "default",
-    backendProfile: source.backendProfile ?? "codex",
-    expiresAt: new Date(Date.now() + 30 * 24 * 60 * 60 * 1000).toISOString(),
-  });
+}
+
+async function sessionRunnerJobBody(args: string[], defaults: Record<string, unknown> = {}): Promise<Record<string, unknown>> {
+  const runnerBody = { ...defaults, ...(await optionalRunnerJsonBody(args)) };
+  copyAgentRunOptions(args, runnerBody, ["image", "namespace", "attempt-id", "runner-id", "source-commit", "service-account-name"]);
+  const managerUrl = agentRunOption(args, "runner-manager-url");
+  if (managerUrl !== null) runnerBody.managerUrl = managerUrl;
+  const runnerIdempotencyKey = agentRunOption(args, "runner-idempotency-key");
+  if (runnerIdempotencyKey !== null) runnerBody.idempotencyKey = runnerIdempotencyKey;
+  return runnerBody;
 }
 
 async function agentRunRestRequest(command: string, method: AgentRunHttpMethod, pathValue: string, body?: unknown): Promise<Record<string, unknown>> {
@@ -3967,12 +4001,26 @@ function jsonInputDisclosureFromArgs(args: string[]): Record<string, unknown> {
   };
 }
 
-function newAgentRunSessionId(): string {
-  return `sess_${Date.now().toString(36)}_${Math.random().toString(36).slice(2, 8)}`;
-}
-
 function defaultAgentRunExecutionPolicy(profile: string): Record<string, unknown> {
-  return { backendProfile: profile, valuesPrinted: false };
+  const keys = profile === "sub2api" || profile === "codex" ? ["auth.json", "config.toml"] : ["auth.json", "config.toml"];
+  return {
+    sandbox: "workspace-write",
+    approval: "never",
+    timeoutMs: 900000,
+    network: "enabled",
+    secretScope: {
+      allowCredentialEcho: false,
+      providerCredentials: [
+        {
+          profile,
+          secretRef: {
+            name: `agentrun-v01-provider-${profile}`,
+            keys,
+          },
+        },
+      ],
+    },
+  };
 }
 
 function safeAgentRunEnvelope(envelope: Record<string, unknown>): Record<string, unknown> {
@@ -4046,7 +4094,7 @@ type AgentRunHttpMethod = "GET" | "POST" | "PUT" | "PATCH" | "DELETE";
 type AgentRunFailureKind = "auth-missing" | "auth-failed" | "agentrun-unreachable" | "schema-mismatch" | "unsupported-version" | "validation-failed";
 
 type AgentRunRestCompatGroup = "queue" | "sessions" | "aipod-specs" | "aipods" | "runs" | "commands" | "runner";
-type AgentRunResourceVerb = "get" | "describe" | "events" | "logs" | "result" | "ack" | "cancel" | "dispatch" | "create" | "apply" | "steer" | "send" | "explain";
+type AgentRunResourceVerb = "get" | "describe" | "events" | "logs" | "result" | "ack" | "cancel" | "dispatch" | "create" | "apply" | "send" | "explain";
 type AgentRunResourceKind = "task" | "run" | "command" | "runnerjob" | "session" | "aipodspec";
 type AgentRunOutputMode = "human" | "wide" | "name" | "json" | "yaml";
 
@@ -4471,7 +4519,7 @@ function unsupported(args: string[]): RenderedCliResult {
     `Error: unsupported AgentRun command: ${command}`,
     "",
     "Supported resource commands:",
-    "  agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send",
+    "  agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send",
     "",
     "Compatibility bridge groups:",
     "  agentrun aipod-specs|queue|runs|commands|runner|sessions",

scripts/src/help.ts

@@ -57,14 +57,14 @@ export function rootHelp(): unknown {
       { command: "commander contract|plan --dry-run|smoke --dry-run|approval request --dry-run", description: "Host Codex commander skeleton contract, no-daemon smoke plan, and dry-run approval preview without live bridges or message sends." },
       { command: "hwlab nodes control-plane|git-mirror|secret --node G14 --lane v03", description: "Manage HWLAB node/lane runtime prerequisites for v0.3+ with the node identity passed as data instead of a command family." },
       { command: "hwlab g14 monitor-prs | hwlab g14 control-plane status|apply|trigger-current|runtime-migration|cleanup-runs|cleanup-released-pvs | hwlab g14 git-mirror status|apply|sync|flush | hwlab g14 tools-image status|build", description: "Start the legacy G14 PR monitor, run bounded v0.2 Tekton/Argo control-plane, manual PipelineRun trigger, runtime migration, CI workspace retention, manual devops-infra git mirror/relay maintenance, or fixed HWLAB CI tools image actions; long confirmed trigger/sync/flush actions return async jobs by default." },
-      { command: "agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send|control-plane|git-mirror", description: "Use AgentRun v0.1 resource primitives with low-noise human output by default; legacy bridge groups remain available for raw compatibility." },
+      { command: "agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send|control-plane|git-mirror", description: "Use AgentRun v0.1 resource primitives with low-noise human output by default; session follow-up uses send only and the server decides internal steer vs turn." },
       { command: "platform-infra sub2api plan|apply|status|validate|codex-pool", description: "Deploy Sub2API in G14 platform-infra, manage the YAML-controlled Codex upstream pool, expose the unified API, and inspect marker sentinel state with low-noise reports without printing API keys." },
       { command: "hwlab cd audit --env dev | hwlab cd status --env dev | hwlab cd apply --env dev --dry-run", description: "Legacy D601 HWLAB DEV CD wrapper kept for explicit old-path diagnostics; current HWLAB rollout uses G14 GitOps." },
       { command: "code-agent-sandbox", description: "Independent Code Agent Sandbox service skeleton for adapter, mode, and credential-boundary diagnostics." },
       { command: "schedule list|get|runs|run|retry-run|delete", description: "Manage backend-core scheduled tasks and run history; schedule run <id> supports --wait-ms N and retry-run reuses the failed run's schedule." },
       { command: "schedule upsert-pgdata-backup [--time HH:MM] [--remote-base /SERVER_DATA/UNIDESK_PG_DATA]", description: "Create or update the daily PGDATA physical backup task that uploads monthly rotated archives to Baidu Netdisk." },
       { command: "codex deploy <commitId> [--provider-id D601] [--timeout-ms N]", description: "Disabled legacy Code Queue deploy path; use the dev-only artifact consumer instead." },
-      { command: "codex submit|steer|resume|queue create|queue merge|move", description: "Frozen legacy Code Queue write commands; use agentrun create/apply/steer/send for new commander work. Historical codex task/tasks/output/read/unread/queues remain available for archive troubleshooting." },
+      { command: "codex submit|steer|resume|queue create|queue merge|move", description: "Frozen legacy Code Queue write commands; use agentrun create/apply/send for new commander work. Historical codex task/tasks/output/read/unread/queues remain available for archive troubleshooting." },
       { command: "codex skills-sync --dry-run [--full]", description: "Inspect the controlled runner skills hostPath lifecycle contract without copying files, restarting services, reading secrets, or mutating live runner paths." },
       { command: "codex execution-plane [--full|--raw]", description: "Read-only D601 native k3s Code Queue execution-plane inspection; compares formal deployments, deprecated Compose residuals, commit markers, pod digest, and mounted worktree HEAD." },
       { command: "codex pr-preflight [--remote] [--push-dry-run --push-dry-run-ref refs/heads/probe/<name>] [--pr-create-dry-run --pr-create-dry-run-head <head>] [--issue N] [--full|--raw]", description: "Read-only PR admission check with compact commander output by default; use --full or --raw to expand the full runtime preflight, tool, and observation payload." },
@@ -75,7 +75,7 @@ export function rootHelp(): unknown {
       { command: "codex read <taskId>", description: "Mark one reviewed terminal task read and return terminal metadata plus final response; prompt/tool logs stay behind drill-down commands." },
       { command: "codex dev-ready", description: "Fetch execution-container readiness, including sanitized skill injection status from /api/dev-ready." },
       { command: "codex judge <taskId> --attempt N [--dry-run] [--include-prompt]", description: "Replay one stored Code Queue attempt through the same judge context builder and MiniMax judge call path used by the live queue worker." },
-      { command: "codex steer <taskId> / codex resume <taskId>", description: "Frozen legacy execution mutation entries; use agentrun steer/send/logs/events/result/ack/cancel against AgentRun resources instead." },
+      { command: "codex steer <taskId> / codex resume <taskId>", description: "Frozen legacy execution mutation entries; use agentrun send/logs/events/result/ack/cancel against AgentRun resources instead." },
       { command: "codex steer-confirm <taskId> --steer-id <id> [--raw]", description: "Read-only lookup for a steerId in task trace so deliveryUnconfirmed can be resolved without resending the corrective prompt." },
       { command: "codex interrupt|cancel <taskId>", description: "Request interrupt for a running Code Queue task, or cancel a queued/retry_wait task, through the same private proxy." },
       { command: "codex queues [--full|--all] [--limit N] [--page N|--offset N]", description: "Read legacy Code Queue archive summaries. Legacy queue create/merge and move are frozen; use agentrun create/apply/get/cancel for new work." },
@@ -391,7 +391,7 @@ function codexHelp(): unknown {
       "bun scripts/cli.ts codex execution-plane [--full|--raw]",
       "bun scripts/cli.ts codex pr-preflight [--remote] [--push-dry-run --push-dry-run-ref refs/heads/probe/<name>] [--pr-create-dry-run --pr-create-dry-run-head <head>] [--issue N] [--full|--raw]",
       "bun scripts/cli.ts codex judge <taskId> --attempt N [--dry-run] [--include-prompt]",
-      "bun scripts/cli.ts agentrun steer session/<sessionId> --prompt-stdin",
+      "bun scripts/cli.ts agentrun send session/<sessionId> --prompt-stdin",
       "bun scripts/cli.ts codex steer <taskId> # frozen legacy write entry",
       "bun scripts/cli.ts codex resume <taskId> # frozen legacy write entry",
       "bun scripts/cli.ts codex steer-confirm <taskId> --steer-id <id> [--raw]",
@@ -463,7 +463,7 @@ function codexHelp(): unknown {
       redline: "data.supervisor.activeRunning.redline names the count field, routine target, burst redline, hard redline, and decisionReady flag.",
       limitSemantics: "filters.requestedLimit preserves the user input; filters.limit/effectiveLimit shows the capped query budget; section outputBudget/rowPage show returned-row caps.",
     },
-    description: "Operate legacy Code Queue as a read-only archive through bounded task/output/read/unread/queues views. New task dispatch, retry/resume, steer, queue mutation, move, and workdir mutation are frozen and replaced by AgentRun resource primitives via bun scripts/cli.ts agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send.",
+    description: "Operate legacy Code Queue as a read-only archive through bounded task/output/read/unread/queues views. New task dispatch, retry/resume, queue mutation, move, and workdir mutation are frozen and replaced by AgentRun resource primitives via bun scripts/cli.ts agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send.",
   };
 }
 
@@ -578,7 +578,7 @@ function artifactRegistryHelp(): unknown {
 
 function agentRunHelpSummary(): unknown {
   return {
-    command: "agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|steer|send|control-plane|git-mirror",
+    command: "agentrun get|describe|events|logs|result|ack|cancel|dispatch|create|apply|send|control-plane|git-mirror",
     output: "human by default; use -o json|yaml or --raw",
     usage: [
       "bun scripts/cli.ts agentrun get tasks --queue commander --limit 20",