Claude-codex-coop
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| health_checkB | 检查 Codex/Claude CLI、工作区、运行目录和写入模式，不启动模型任务。 |
| get_model_catalogB | 识别主对话 AI，并读取另一端协作 AI 的候选模型、推理强度和已保存设置。 |
| show_partner_selectorA | 在当前对话中显示一个紧凑的协作 AI 模型与推理强度选择条；选择会保存并用于之后的 start_workflow 调用。 |
| open_sidebarA | 准备当前工作区的本地协作控制页（显示协作伙伴模型、推理强度与运行状态），返回 URL。Codex 用 open_in_codex 在 right 面板打开；Claude 桌面版用内置浏览器面板 preview_start 打开。此工具本身不会打开面板。 |
| set_collaboration_modeA | 打开后，当前宿主中的后续实质任务会自动交给另一端 AI 协作；用户无需再调用 workflow。 |
| set_partner_preferencesB | 保存另一端 AI 的自动/手动选模、模型 ID 和推理强度，供之后的原生聊天协作调用使用。 |
| start_workflowC | 当前聊天 AI 自动担任主协调者，把本轮任务交给另一端 AI；完成后主 AI读取结果、继续追问或向用户汇总。 |
| start_consultationC | 后台启动一次 Codex 或 Claude 的独立只读咨询。 |
| start_model_debateB | 后台启动 Claude 与 Codex 的独立方案、交叉质询和证据裁决；不会修改项目。 |
| start_codex_implementationB | 按用户已批准的决策让 Codex 修改项目并验证。服务需显式启用 --allow-write。 |
| get_run_statusB | 查询后台协作任务状态；完成后返回最终报告。 |
| list_runsB | 列出最近的协作运行记录。 |
| get_usage_summaryA | 读取 ai-coop 单次或最近多次任务的 Codex/Claude token 与费用统计；不代表账户剩余额度。 |
| project_statusA | 读取项目 Git 状态和差异统计；非 Git 目录则列出顶层文件。 |
| read_project_fileB | 读取所选工作区内的 UTF-8 文本文件，禁止越界访问。 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| open_ai_coop | 检查 AI Coop 连接，并在当前对话显示协作模型与推理强度选择。 |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| ai-coop-partner-selector | 紧凑显示另一端 AI 的模型与推理强度选择。 |
TDQS
Scored across 15 tools
The start_* tools represent distinct workflows (workflow, consultation, debate, implementation), and the setup/status tools are mostly distinct. However, set_collaboration_mode and start_workflow both concern handing off work to the partner AI, and start_consultation vs start_model_debate could be confused without reading descriptions carefully.
Tool names are consistently snake_case and mostly follow a verb_noun pattern such as get_model_catalog, set_partner_preferences, and start_workflow. A few names break the pattern (health_check, project_status) but overall the convention is predictable.
At 15 tools, the set is within the expected range for a collaboration-orchestration server and each tool appears to serve a distinct role in setup, execution, or monitoring. No obvious redundant or filler tools are present.
The surface covers setup, model selection, workflow starts, status polling, run listing, usage summaries, and implementation runs. A notable gap is the lack of an explicit cancel/stop tool for background collaboration runs, which could leave agents unable to abort a running task.