范围： 仅限本地服务器或你拥有/获准使用的服务器。不是用于公共服务器自动化的工具。
MCP on My SAMP 让 AI 智能体能够以可重复的工作流来测试游戏服务器：
启动 / 停止 / 检查 open.mp 的状态；
启动 / 停止 / 检查无头 RakClient 的状态；
等待客户端真正进入 Spawned 状态；
发送允许的斜杠命令；
读取已缓冲的客户端输出历史；
验证客户端是否已收到服务器的响应；
从 Pawn 游戏模式的源码中发现命令；
拒绝允许列表之外的命令。
它不自带 flood、垃圾信息、延迟注入、任意 RCON，也不对公共服务器做自动化。
Related MCP server: Brainstorm
flowchart LR
A[AI Agent] -->|MCP stdio| B[MCP on My SAMP]
B --> C[open.mp Server]
B --> D[Headless RakClient]
D -->|UDP localhost| C
C -->|server response| D
D --> B
B -->|assertion| A有效的往返（round-trip）证据：
command dikirim
→ server callback menerima command
→ gamemode mengirim response
→ client menerima response
→ MCP assertion berhasil仅凭 Spawned 并不能证明命令成功执行。
需要 Python 3.10 或更高版本。
在仓库根目录下运行：
Windows
py -3 -m venv .venv
.venv\Scripts\activate
python -m pip install --upgrade pip
python -m pip install ".[dev]"Linux / macOS
python3 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install ".[dev]"pytest -q预期输出：
35 passedopen.mp 和 RakClient 的二进制文件仅用于实时测试。Python 单元测试不需要这些二进制文件也可以运行。
从模板创建本地配置文件：
Windows
copy config.example.json local-server.jsonLinux / macOS
cp config.example.json local-server.json填写文件内容：
{
"executable": "vendor/openmp/Server/omp-server.exe",
"working_dir": "vendor/openmp/Server",
"args": ["--config-path", "config.json"],
"ready_text": "Legacy Network started on port",
"startup_timeout": 30
}根据自己电脑上 open.mp 的位置修改 executable 和 working_dir。local-server.json 不会提交到 Git，因为不同电脑的路径不同。
mcp-gta-samp --config local-server.jsonWindows
mcp-gta-samp ^
--config local-server.json ^
--client-executable vendor/rakclient-bin/rakclient.exe ^
--client-arg --server ^
--client-arg 127.0.0.1:7777 ^
--client-arg --nick ^
--client-arg MCPBot ^
--client-arg --scripts-dir ^
--client-arg vendor/rakclient-bin/scripts ^
--gamemode-source vendor/openmp/Server/gamemodes/mcp_test.pwnLinux / macOS
mcp-gta-samp \
--config local-server.json \
--client-executable vendor/rakclient-bin/rakclient \
--client-arg --server \
--client-arg 127.0.0.1:7777 \
--client-arg --nick \
--client-arg MCPBot \
--client-arg --scripts-dir \
--client-arg vendor/rakclient-bin/scripts \
--gamemode-source vendor/openmp/Server/gamemodes/mcp_test.pwnMCP 使用 stdio 传输。
工具 | 功能 |
| 启动服务器并等待就绪。 |
| 检查服务器状态和 PID。 |
| 停止服务器。 |
| 启动无头 RakClient。 |
| 检查客户端状态。 |
| 停止客户端。 |
| 在进入 |
| 获取已缓冲的客户端输出。 |
| 断言客户端已收到特定输出。 |
| 从 Pawn 源码中显示命令。 |
| 针对允许列表验证命令。 |
最后两个工具仅在指定 --gamemode-source 时可用。
1. server_status
2. server_start jika belum berjalan
3. client_start
4. tunggu state Spawned
5. server_list_commands
6. server_assert_command("/help")
7. client_send_chat("/help")
8. client_assert_output("MCP Test Commands:")
9. client_get_history bila perlu diagnosis
10. client_stop
11. server_stop给智能体的重要指令：
不要访问公共服务器；
不要将 boot、join 或 Spawned 视为命令完成的往返；
如果失败，请按边界分类：boot、连接、spawn、队列、出站数据包、服务器回调、服务器响应、客户端解析器，或 MCP 断言；
测试结束后一律停止进程；
确保 UDP 端口 7777 重新恢复闲置。
测试游戏模式位于：
vendor/openmp/Server/gamemodes/mcp_test.pwn可用的命令：
/help
/status如果修改 Pawn 源码，请从服务器目录重新编译 .amx：
qawno\pawncc.exe -i.\qawno\include -o.\gamemodes\mcp_test .\gamemodes\mcp_test.pwn实时工作流程：
server_start
→ client_start
→ client mencapai Spawned
→ client_send_chat("/help")
→ client_assert_output("MCP Test Commands:")
→ client_assert_output("/status - show a test response")
→ client_stop
→ server_stop无头 RakClient 可以证明协议、状态和命令是否正常。它不会生成截图。视觉测试需要单独的、带渲染的 GTA 客户端。
运行测试：
pytest -q构建 wheel：
python -m pip wheel . --no-deps -w dist安装 wheel：
python -m pip install dist/mcp_gta_samp-0.1.0-py3-none-any.whl核心结构：
mcp_gta_samp/ package Python dan MCP facade
tests/ unit, contract, dan bridge tests
config.example.json template konfigurasi
vendor/ binary dan fixture live test此 MCP 将使用范围限制在本地/自有服务器。请不要把凭据、代理池、私有配置、私有日志或服务器测试数据放进公共仓库。
如果服务器需要公网访问，请自行添加身份验证和网络隔离。本包不是为公开的游戏控制 API 而设计的。
MIT License。请查看 LICENSE。
问题反馈：报告问题
Start the configured local open.mp test server and wait for readiness.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already cover idempotency (idempotentHint=true), read-only status (readOnlyHint=false), and destructiveness (destructiveHint=false). The description adds the 'wait for readiness' behavior, which is beyond what annotations provide, giving the agent a clear expectation of the tool's runtime behavior.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, front-loaded sentence with no filler. Every word contributes meaning: 'start', 'configured local open.mp test server', and 'wait for readiness' are all necessary and sufficient.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
This is a simple tool with no parameters, an output schema, and annotations covering idempotency and safety. The description fully covers the intended action and readiness behavior, leaving no critical gaps for an agent to call it correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has zero parameters, so there are no parameter semantics to document. Per the baseline for 0 params, a score of 4 is appropriate; the description doesn't need to compensate for schema gaps.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description states a specific verb ('Start') and resource ('configured local open.mp test server') and adds 'wait for readiness' as a behavioral outcome. This clearly distinguishes it from siblings server_stop and server_status through action and intent.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The start action and readiness wait make the context obvious: this is for bringing the server up. It does not explicitly name alternatives or when-not conditions, but the sibling names and the verb provide clear contextual guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return local open.mp test server process status.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The annotations already declare readOnlyHint=true and destructiveHint=false, so the safe read-only nature is covered. The description adds the 'local' and 'process status' scope, but does not provide additional behavioral detail such as what exactly the status output contains or whether it might throw errors if no server is running. With the output schema present, the agent can likely infer return structure, so the description adds minimal but acceptable value.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single, front-loaded sentence delivers the essential meaning with no filler. It names the resource and the type of operation without unnecessary elaboration.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a no-parameter, read-only status tool with an output schema present, the description is complete enough to guide correct invocation. There are no prerequisites, options, or alternatives that need clarifying beyond what the summary already provides.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
There are zero parameters, so the description has no parameter meaning to add. The baseline for no-parameter tools is 4, and there is no relevant gap to compensate for.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses a specific verb ('Return') and a clear resource ('local open.mp test server process status'), making the tool's function immediately obvious. It also differentiates itself from siblings server_start and server_stop by describing a read-only status check rather than a state-changing operation.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description clearly implies the tool is for checking status rather than starting or stopping the server, which are the sibling operations. It does not explicitly state 'use this when you need to check status and not start/stop', but the context and naming make the intended use clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Stop the configured local open.mp test server.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
No output parameters | ||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare destructiveHint=true and idempotentHint=false, but the description adds no behavioral context beyond the verb 'Stop'. It does not clarify consequences such as terminating active sessions or clearing server state. There is no contradiction with annotations, but the description relies entirely on structured metadata.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, front-loaded sentence with no filler. It names the action and the target directly, and every word earns its place.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a zero-parameter destructive action with annotations covering safety and an output schema present, the description is sufficient for basic invocation. It is slightly thin on operational context such as behavior when the server is already stopped, but that is not necessary for calling the tool correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool accepts zero parameters and schema description coverage is 100%, so there are no parameter semantics to document. The baseline of 4 applies because the description cannot add meaning beyond the empty schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description states a precise action ('Stop') and the exact target ('configured local open.mp test server'). This clearly distinguishes it from sibling tools server_start and server_status, which have complementary purposes.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is given on when to use this tool versus server_start or server_status. The description only states the action and target, leaving the agent to infer usage from the tool name and siblings rather than from explicit instructions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0server_startserver_statusserver_stopScored across 3 tools
Each tool has a single, clear responsibility: start, stop, or check status. There is no overlap or ambiguity between them.
All tools follow the identical pattern of 'server_' followed by a simple verb (start, stop, status). This is perfectly consistent and predictable.
Three tools is exactly right for the narrow scope of managing a local server. Each tool is essential and earns its place without unnecessary bloat.
The tool set covers the core lifecycle (start, stop, status) but lacks a restart operation, which is a common need. However, agents can still accomplish restart via stop+start, so this is a minor gap.
Create guides as MCP servers to instruct coding agents to use your software (library, API, etc).
Official MCP server for Agentwork — delegate tasks to AI agents with human-in-the-loop
Hosted MCP server connecting claude.ai, ChatGPT and other AI apps to your own computer
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
