codex-app-server

0 Fork 0

Files

T

host 1c8fc3f7c8 对齐 Muse P4：浏览器完全对齐（CDP broker 单元外 + a11y-tree-only + 确定性分类器）

- chv-guest/chv-cdp-broker.py（新）：单元外 CDP broker，用户 chv-browser，
  独家 chromium.launch(headless=True, args=LAUNCH_ARGS)；启动参数白名单
  无 --remote-debugging-port、无 devtools；导航前 URL 校验 + 拦截名单；
  快照恒带 untrusted_input 与 injection_flags；高风险表单在页时 open
  即 ask 入队；观测到导航 POST 自动暂停；命中拦截规则的下载取消 + 暂停
  + ask；全部回包与错误经 sanitize_error + response_leaks 检查。
- chv-guest/chv-browser.py（P2 版重构）：改为 broker 客户端，不再直连
  Playwright（本文件无 playwright import），保留 URL 校验 + 快照截断
  作为第二层防线；暂停状态与 broker 同步。
- chv-guest/chv_browser_guard.py（新）：纯 stdlib 共享规则——validate_url、
  拦截名单加载/匹配、快照截断、错误消毒（去 HTML、截断 200 字符、凭据形
  key=value 与 sk- 形 token 打码）、注入启发式（中英文各 6 条）、表单风险
  扫描、下载扫描（可执行扩展名）、文件型 ask 队列（pending/allowed/denied，
  与 P3 ApprovalQueue 同语义）、启动参数与回包泄露检查、脚本注入面
  禁止清单（FORBIDDEN_PAGE_ATTRS/FORBIDDEN_OPS）。
- chv-guest/browser-blocklist.txt（新）：静态拦截名单；首版只有 RFC 2606
  .test 自测条目，生产条目由运维构建时填入，无在线更新。
- chv-guest/build.sh/buildinfo.py：P2 三脚本 + P4 三文件纳入镜像与
  buildinfo；新增 chv-cdp-broker.service 与 /run/chv/ask-queue tmpfiles。
- docs/chv-muse-p4-gap.md（新）：现状 vs Muse 官方要求差距表、改动、
  诚实边界、P2 合入约定、真机待验证清单（round 12）。
- docs/chv-production-readiness.md：补 §15（P4 一节；编号合入时若与
  P2 的 §15 撞，顺延）。
- cloud/tests/test_chv_muse_p4.py（新，36 项）：guard 单元、AST 钉住
  禁止属性、假 Playwright broker 测试、假 broker agent 测试、真 Unix
  socket 联调、buildinfo/build 干跑断言；含否定测试（evaluate 被拒、
  DOM 外泄被拒、DevTools 不可达、私有地址被拒、高风险表单被拦、
  注入文本被标注）。

测试：P4 新测试 36/36；全量 cloud 套件 1903 项，两次连续全过
（skipped=8）。附记：第一次全量跑出 1 个失败但输出未保留、后两次
均未复现，疑为时序抖动，未追认。
真机验证未做：排队等 round-11/第三棒释放 100.65.51.33，清单见 gap
表 §5。本地提交，未推未合未部署。

2026-11-06 22:36:56 +00:00

test_api_keys.py

多 key：OAuth 登录签发附加 key 不再轮换，keys 管理端点

2026-10-26 16:27:42 +08:00

test_app_token.py

App-token：推分支建 PR 改走 GitHub App installation token

2026-10-26 16:47:55 +08:00

test_approvals_server.py

feat: P1 补齐（自动评审队列、自动化+分诊、并发 caps、运行时审批队列）

2026-10-27 15:53:50 +08:00

test_approvals.py

feat: P1 补齐（自动评审队列、自动化+分诊、并发 caps、运行时审批队列）

2026-10-27 15:53:50 +08:00

test_auth.py

fix(auth): unknown project role denies instead of KeyError 500

2026-10-30 00:49:03 +08:00

test_auto_memory.py

自动记忆：官方 memories 对齐的提取+注入闭环（确定性 v1）

2026-10-28 12:12:10 +08:00

test_automation_run_due.py

fix(automations): 审计行不再显式插 TEXT id，PG identity 下不再丢审计

2026-10-30 01:55:59 +08:00

test_automations_server.py

feat: P1 补齐（自动评审队列、自动化+分诊、并发 caps、运行时审批队列）

2026-10-27 15:53:50 +08:00

test_automations.py

feat: P1 补齐（自动评审队列、自动化+分诊、并发 caps、运行时审批队列）

2026-10-27 15:53:50 +08:00

test_autoscaler.py

autoscaler: guard scale-down against live coordinators and busy victims

2026-10-29 00:19:30 +08:00

test_caps_server.py

P1 security batch: quota execution checks, cost accounting, goal/coordinator semantics, PR channel, orphan signature, control-plane and console hardening

2026-10-27 16:27:50 +00:00

test_caps.py

feat: P1 补齐（自动评审队列、自动化+分诊、并发 caps、运行时审批队列）

2026-10-27 15:53:50 +08:00

test_chat_history.py

chat: 持久化继续消息，chat-history 交替历史

2026-10-29 07:01:57 +08:00

test_chatgpt_codex_skills.py

feat: add software-engineering skill v0.2

2026-10-30 06:00:31 +00:00

test_child_model.py

feat: 子任务默认模型/强度（child_model/child_effort）

2026-10-29 05:27:22 +08:00

test_chv_arch95.py

fix(chv): CHV 54 VmConfig 形态——payload.kernel/cmdline 纯字符串 + 磁盘显式 image_type Raw

2026-11-06 17:59:53 +00:00

test_chv_arch95plus.py

feat(chv): 架构/可维护性/前端只读侧推进——生产实验分离、配置片段化、VM 只读展示

2026-10-31 06:55:26 +00:00

test_chv_arch_maint_fe.py

feat(chv): 架构/可维护性/前端只读侧推进——生产实验分离、配置片段化、VM 只读展示

2026-10-31 06:55:26 +00:00

test_chv_closeup.py

fix(chv): restore 路径补 H7 net 校验+登记，seed 暂存文件走 atomic_file

2026-11-06 19:32:41 +00:00

test_chv_cover100.py

fix(chv): api_client 接受 204 无 Content-Length 响应

2026-11-06 17:43:54 +00:00

test_chv_fix.py

fix(chv): CHV 54 VmConfig 形态——payload.kernel/cmdline 纯字符串 + 磁盘显式 image_type Raw

2026-11-06 17:59:53 +00:00

test_chv_gaps.py

fix(chv): restore 路径补 H7 net 校验+登记，seed 暂存文件走 atomic_file

2026-11-06 19:32:41 +00:00

test_chv_gitea_sync.py

feat(chv-guest): seed 附加项供客机用户与静态网络（wake 路径前置）

2026-11-06 16:58:27 +00:00

test_chv_muse_p4.py

对齐 Muse P4：浏览器完全对齐（CDP broker 单元外 + a11y-tree-only + 确定性分类器）

2026-11-06 22:36:56 +00:00

test_chv_optimize.py

feat(chv): 架构/可维护性/前端只读侧推进——生产实验分离、配置片段化、VM 只读展示

2026-10-31 06:55:26 +00:00

test_chv_residual.py

fix(chv): P2-20/P2-19/P2-17/P2-18 残留收尾 + NO_PROXY 合并

2026-11-06 18:41:50 +00:00

test_chv_secfix.py

test(chv-guest): 构建脚本测试夹具跟进 B-09 新契约

2026-11-06 15:34:06 +08:00

test_chv_to10.py

fix(chv-guest): 客机镜像构建缺口补齐（B-02/B-03/B-04/B-05/B-06/B-07/B-08/B-09/B-10/B-13/B-15）

2026-11-06 15:13:23 +08:00

test_cloud_dispatch_pr.py

P1 security batch: quota execution checks, cost accounting, goal/coordinator semantics, PR channel, orphan signature, control-plane and console hardening

2026-10-27 16:27:50 +00:00

test_cloud_hypervisor.py

fix(chv): CHV 54 VmConfig 形态——payload.kernel/cmdline 纯字符串 + 磁盘显式 image_type Raw

2026-11-06 17:59:53 +00:00

test_console.py

console: React 生产级重写（左列表 + 中对话 + 右详情）

2026-10-29 06:32:38 +08:00

test_coord_quota_gate.py

feat: 跨 worker 真 kill（kill_requests 意图表 + 收殓器）与自动化触发入闸

2026-10-30 01:32:35 +08:00

test_coordinator_row_vanished.py

协调员行消失按中断处理：本体行消失即放槽，子任务行消失按 interrupted 合流

2026-10-28 08:57:05 +00:00

test_coordinator.py

worker: release running slot after coordinator completes

2026-10-29 02:42:52 +08:00

test_cost_files_markers.py

成本配额与文件只读标记：task-cost/quota-status/task-files

2026-10-30 05:11:06 +08:00

test_db_backend.py

Postgres real backend: psycopg adapter, schema, migration

2026-10-26 03:23:06 +08:00

test_db.py

cloud: fill gaps (memory/KB/cost/models endpoints, WAL concurrency, OAuth scaffold, push subscriptions)

2026-10-26 01:32:57 +08:00

test_docker_runtime.py

Docker tasks: --docker-runtime option (runsc support, default off)

2026-10-26 10:26:21 +08:00

test_docker_sandbox.py

feat: Docker 沙箱 /tmp tmpfs 隔离 + 可选根只读（内层 bwrap 本体不在本仓修复范围）

2026-10-30 04:17:09 +08:00

test_fallback_retry.py

Perf: light worker poll query + tasks(status) index

2026-10-26 06:19:01 +08:00

test_github_app.py

feat: GitHub App manifest 全量获取：仓库/组织/企业/账号四类最高档

2026-10-30 02:53:46 +08:00

test_github_repos.py

P1 security batch: quota execution checks, cost accounting, goal/coordinator semantics, PR channel, orphan signature, control-plane and console hardening

2026-10-27 16:27:50 +00:00

test_goal_mode.py

console: React 生产级重写（左列表 + 中对话 + 右详情）

2026-10-29 06:32:38 +08:00

test_goal_resume.py

worker: auto-resume falls back to fresh goal for never-started tasks

2026-10-29 16:45:07 +08:00

test_handler_500.py

fix: 请求处理未捕获异常回 500 JSON，不再断连变 nginx 502

2026-10-27 10:56:08 +08:00

test_harness2_tasks.py

harness2: per-role models, sweep batch, task file assets endpoint

2026-10-29 16:32:58 +08:00

test_harness2.py

harness2: recursive harness optimization wired to real codex tasks

2026-10-29 14:21:18 +08:00

test_host_sandbox_routing.py

Fix codex-host profile: capability allow + netlink (bwrap needs both)

2026-10-26 08:39:57 +08:00

test_identity_memory.py

云端身份与记忆：用户画像+分级记忆，不动 thread

2026-10-28 03:43:02 +08:00

test_imagegen_worker.py

feat: 图片生成 MCP 工具 image_generate（中转站图片分组）

2026-10-30 02:53:30 +08:00

test_imagegen.py

feat: 图片改图 MCP 工具 image_edit

2026-10-30 04:05:00 +08:00

test_kill_requests.py

feat: 跨 worker 真 kill（kill_requests 意图表 + 收殓器）与自动化触发入闸

2026-10-30 01:32:35 +08:00

test_longtask_auto_resume.py

feat: auto-resume platform-interrupted non-goal tasks

2026-10-30 15:39:55 +08:00

test_mainline_goal_command.py

feat: 不用斜杠，单发 goal 也触发 /goal 目标

2026-10-30 07:15:05 +08:00

test_marker_tools.py

任务内标记工具：task-create/task-poll/task-read 移植、wait-threads 接通、新增 task-cancel/task-retry

2026-10-28 10:13:40 +00:00

test_marker_wait_heartbeat.py

worker: heartbeat parent task during marker poll/wait loops

2026-10-29 04:40:47 +08:00

test_mcp_server.py

feat: project directory tree + first-round snapshot

2026-10-30 14:46:11 +08:00

test_member_delete.py

残留批次A：成员删除、PR needs_local_worker 落库、推送 DNS TOCTOU

2026-10-27 19:10:49 +00:00

test_memory_kb.py

cloud: production control plane (API, auth/RBAC, queue, sandbox, memory/KB, fallback/cost, webhooks)

2026-10-26 01:23:21 +08:00

test_models_cost.py

feat: 任务详情 LLM 请求数（cost_records 存 completions）

2026-10-30 12:58:41 +08:00

test_multiuser_gaps.py

P0 security batch: true global admin, quota gates on all spending paths, default quota for quota-less users, fix-review author trust + docker/owner attribution, GitHub App setup issuer binding, real cancel kill, wider shell env exclusion

2026-10-27 15:35:23 +00:00

test_neverstart_fresh.py

worker: auto-resume falls back to fresh goal for never-started tasks

2026-10-29 16:45:07 +08:00

test_oauth_bind.py

多 key：OAuth 登录签发附加 key 不再轮换，keys 管理端点

2026-10-26 16:27:42 +08:00

test_oauth_real_login.py

多 key：OAuth 登录签发附加 key 不再轮换，keys 管理端点

2026-10-26 16:27:42 +08:00

test_oauth.py

cloud: fill gaps (memory/KB/cost/models endpoints, WAL concurrency, OAuth scaffold, push subscriptions)

2026-10-26 01:32:57 +08:00

test_ops_batch.py

HA stage 0: health reads workers registry, daily pg_dump timer, backup freshness in monitor (#48 )

2026-10-26 23:51:54 +08:00

test_outer_hardening.py

Outer hardening: AppArmor profile, pids limit, egress allowlist proxy

2026-10-26 07:42:14 +08:00

test_p0_security.py

2026-10-27 15:35:23 +00:00

test_p0_three.py

断点镜像状态标签回落：host-state 条目写 status 不是 goal_status

2026-10-29 15:31:07 +08:00

test_p1_security_server.py

P1 security batch: quota execution checks, cost accounting, goal/coordinator semantics, PR channel, orphan signature, control-plane and console hardening

2026-10-27 16:27:50 +00:00

test_p1_security_worker.py

fix: Docker 断点续跑透传密钥环境（SUB2API_API_KEY 缺失空转修法）

2026-10-29 12:42:19 +08:00

test_p2_console.py

console: React 生产级重写（左列表 + 中对话 + 右详情）

2026-10-29 06:32:38 +08:00

test_p2_migrations.py

P2-ops: migration alignment, deploy test list, monitor alerts/single-host, backup/rollback closure, Caddy template

2026-10-27 16:37:28 +00:00

test_peer_messages_server.py

feat: 任务间互发消息（API + agent 标记块，不中断任务）

2026-10-27 14:12:28 +08:00

test_peer_messages.py

feat: 任务间互发消息（API + agent 标记块，不中断任务）

2026-10-27 14:12:28 +08:00

test_playwright_bridge.py

perf: 桥变化才写、stdin 优先、stream 50ms 轮询

2026-10-30 13:36:24 +08:00

test_playwright.py

perf: Chromium 启动 flag 微调 + screencast 画质 q60->q50

2026-10-30 14:27:04 +08:00

test_postgres_backend.py

fix(db): register auto-memory tables in _OR_REPLACE_PK

2026-10-28 22:49:06 +08:00

test_postgres_integration.py

Postgres real backend: psycopg adapter, schema, migration

2026-10-26 03:23:06 +08:00

test_pr_app_token.py

App-token：推分支建 PR 改走 GitHub App installation token

2026-10-26 16:47:55 +08:00

test_pr_commit.py

pr: commit worktree changes before push; fix merge/pr order docs

2026-10-26 15:20:02 +08:00

test_pr_jobs.py

feat: G7 分层 fail-open/fail-closed + G8 Docker 轮询掐断 + G9 worker 忽略 host 预算 + G13/G14 归属（第三批）

2026-10-28 11:14:10 +08:00

test_pr_needs_local.py

残留批次A：成员删除、PR needs_local_worker 落库、推送 DNS TOCTOU

2026-10-27 19:10:49 +00:00

test_pr_timeout.py

pr 超时：git/gh 子进程默认 120 秒超时，超时转 502 不挂 handler

2026-10-26 17:01:47 +08:00

test_prestart_stages.py

worker: cover goal docker paths with prestart stage logging

2026-10-29 17:46:12 +08:00

test_prod_gaps.py

feat: /metrics 扩队列深度/worker 聚合/分模型成本/未定价占比/近24h失败

2026-10-30 04:15:54 +08:00

test_prod_webhook_exec.py

429 限流加 Retry-After 头：值为限流窗口秒数，不改阈值与计数

2026-10-28 10:12:57 +00:00

test_prod_worker_timeout.py

feat: 长任务补齐——流式进度 + 跨进程中断 + worker 超时约定

2026-10-27 03:05:15 +08:00

test_project_repo.py

feat: 仓库接入通用化——批量预检 + 多仓 ssh 别名配给

2026-10-30 04:18:09 +08:00

test_project_todos_server.py

待办：项目待办模型 + /projects/{id}/todos 接口 + 控制台待办页

2026-10-30 05:11:06 +08:00

test_project_todos.py

待办：项目待办模型 + /projects/{id}/todos 接口 + 控制台待办页

2026-10-30 05:11:06 +08:00

test_push_toctou.py

残留批次A：成员删除、PR needs_local_worker 落库、推送 DNS TOCTOU

2026-10-27 19:10:49 +00:00

test_push.py

feat: goal 断点续跑端点（G10）+ 摘要持久化与推送 outbox（G11/G12）

2026-10-28 11:09:29 +08:00

test_queue_sandbox.py

fix(worker): 无 result_text_full 列的老库/手工表写回回落，不拖垮终局

2026-10-30 02:46:37 +08:00

test_rate_limit_db.py

fix: 限流清理移出请求路径，根治 rate_limit_hits 死锁

2026-10-27 05:40:30 +08:00

test_relay.py

feat: per-task model/effort/approval + relay models endpoint (#59 )

2026-10-27 01:44:53 +08:00

test_review_server.py

feat: P1 补齐（自动评审队列、自动化+分诊、并发 caps、运行时审批队列）

2026-10-27 15:53:50 +08:00

test_review.py

feat: P1 补齐（自动评审队列、自动化+分诊、并发 caps、运行时审批队列）

2026-10-27 15:53:50 +08:00

test_run_all_pool.py

worker: always start capacity pool threads in run_all

2026-10-29 05:41:35 +08:00

test_server_browser_proxy.py

feat: browser stream waits through round gaps instead of disconnecting

2026-10-30 15:14:37 +08:00

test_server_endpoints.py

修项目成本按天接口在 Postgres 上的 Decimal 序列化崩溃

2026-10-26 23:17:25 +08:00

test_server_goal_notify.py

P1 security batch: quota execution checks, cost accounting, goal/coordinator semantics, PR channel, orphan signature, control-plane and console hardening

2026-10-27 16:27:50 +00:00

test_server_merge.py

Docker sandbox real test + cloud merge status endpoint

2026-10-26 02:46:10 +08:00

test_server_oauth_push.py

2026-10-27 15:35:23 +00:00

test_server_pr_app_token.py

App-token：推分支建 PR 改走 GitHub App installation token

2026-10-26 16:47:55 +08:00

test_server_retry.py

Worker concurrency (capacity enforced), survival (heartbeat/sweep/drain), retry endpoint

2026-10-26 06:46:20 +08:00

test_server_task_files.py

fix: task files endpoint resolves tenant docker worktrees

2026-10-29 23:16:01 +08:00

test_server_task_tree.py

feat: project directory tree + first-round snapshot

2026-10-30 14:46:11 +08:00

test_server.py

2026-10-27 15:35:23 +00:00

test_session_goal_automation.py

feat: 会话内 LLM 自判 goal / 自动化标记工具，不靠关键词

2026-10-30 07:15:05 +08:00

test_shared_home_mcp.py

feat: sync shared-home MCP scripts on control-plane startup

2026-10-30 07:08:51 +00:00

test_skill_seed_worker.py

feat: add software-engineering skill v0.2

2026-10-30 06:00:31 +00:00

test_skills_mcp.py

feat: 线程操作与 skills/MCP 管理面暴露（host CLI + 服务端 API + 控制台）

2026-10-27 16:37:28 +00:00

test_software_engineering_github.py

feat: software-engineering skill GitHub guarantees

2026-10-30 06:33:25 +00:00

test_sub2api.py

fix: 登录限流下开通突发不再硬失败（节流 + 429 重试）

2026-10-27 23:09:22 +08:00

test_task_attachments.py

feat: 任务附件——server/worker/console 链路打通（+ popen goal 新语义补漏）

2026-10-27 04:30:43 +08:00

test_task_continue.py

feat: 旧任务继续会话——终端任务带新消息 + 全部选项同一会话续跑

2026-10-27 04:40:11 +08:00

test_task_environments.py

console: React 生产级重写（左列表 + 中对话 + 右详情）

2026-10-29 06:32:38 +08:00

test_task_links.py

feat: 计划模式 + 后续消息 + 会话分支接云端 (#65 )

2026-10-27 03:17:51 +08:00

test_task_list_markers.py

任务列表 task-list 标记 + 首轮项目上下文注入

2026-10-30 05:11:06 +08:00

test_task_list_pagination.py

console: 任务列表服务端分页 + 控制台分页面签

2026-10-30 02:25:48 +08:00

test_task_list_tokens.py

console: 任务详情也显示首字、总耗时、token

2026-10-30 00:39:13 +08:00

test_task_options.py

feat: G7 分层 fail-open/fail-closed + G8 Docker 轮询掐断 + G9 worker 忽略 host 预算 + G13/G14 归属（第三批）

2026-10-28 11:14:10 +08:00

test_task_quotas.py

test: G14/open 旧口径测试按 P1 口径更新

2026-10-28 11:25:17 +08:00

test_task_trace.py

fix: 租户 Docker 任务链路可见（租户目录优先、共享目录兜底）

2026-10-27 12:58:24 +08:00

test_task_turn_ids.py

console: React 生产级重写（左列表 + 中对话 + 右详情）

2026-10-29 06:32:38 +08:00

test_tenant_isolation.py

Multi-tenant isolation for Docker tasks (per-project mounts + networks + drop caps)

2026-10-26 07:24:54 +08:00

test_thread_lifecycle.py

feat: 线程生命周期对齐 15 项（bindings/refs/快照/心跳绑定/移交落点）

2026-10-28 04:56:22 +08:00

test_thread_ops.py

feat: 线程操作与 skills/MCP 管理面暴露（host CLI + 服务端 API + 控制台）

2026-10-27 16:37:28 +00:00

test_thread_resolve.py

fix: 线程解析按 home×provider 穷尽回退（共享 home + sub2api）

2026-10-28 05:36:18 +08:00

test_thread_unarchive.py

fix: thread/unarchive 走已归档解析（--archived 列表）

2026-10-28 06:10:02 +08:00

test_thread_wait_ops.py

feat: 线程等待与状态读取 8 项（wait/read/message/handoff/worktree/pr-checks/automation/sleep）

2026-10-28 04:33:20 +08:00

test_todo_markers.py

待办：todo-list/create/update 标记工具 + 首轮未完成待办注入

2026-10-30 05:11:06 +08:00

test_tree.py

feat: project directory tree + first-round snapshot

2026-10-30 14:46:11 +08:00

test_webhooks.py

cloud: production control plane (API, auth/RBAC, queue, sandbox, memory/KB, fallback/cost, webhooks)

2026-10-26 01:23:21 +08:00

test_websearch_mcp_env.py

websearch: MCP env 写进租户 config，不只靠 docker -e

2026-10-30 01:16:49 +08:00

test_websearch_mcp.py

feat: web search MCP v1 (oh-my-pi pattern, Python stdlib)

2026-10-28 17:14:22 +08:00

test_websearch_searxng.py

websearch: SearXNG 默认只查 bing,360search

2026-10-30 01:16:49 +08:00

test_websearch_worker.py

fix: pass search env into Docker task containers (SEARXNG_URL was lost)

2026-10-28 22:17:19 +08:00

test_worker_concurrency.py

Worker concurrency (capacity enforced), survival (heartbeat/sweep/drain), retry endpoint

2026-10-26 06:46:20 +08:00

test_worker_peer.py

feat: 任务间互发消息（API + agent 标记块，不中断任务）

2026-10-27 14:12:28 +08:00

test_worker_quota.py

test: G14/open 旧口径测试按 P1 口径更新

2026-10-28 11:25:17 +08:00

test_worker_tree_injection.py

feat: project directory tree + first-round snapshot

2026-10-30 14:46:11 +08:00

test_worker.py

fix(worker): 终态审计带上 error，复发卡死可定位

2026-10-30 01:38:01 +08:00