Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@kscc-mcp-server Fix the flaky test in checkout and refactor the cart module"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
把 kscc CLI 封装成 stdio MCP server，暴露两个工具，供 MCP 客户端（DSH、Claude Desktop、Cursor 等）把编码任务委派给 kscc。
设计文档：
docs/superpowers/specs/2026-09-07-kscc-mcp-server-design.md
run_kscc_prompt / continue_kscc_session 默认 permissionMode=bypassPermissions，即 kscc 不经确认即可执行文件写入 / 命令。仅在受信任环境与受信任调用方使用。 收紧方式：
permissionMode 设为 acceptEdits（自动接受编辑但不自动跑命令）或 plan（只规划）。
allowedTools 限定工具白名单（如 ["Read","Edit","Bash"]）。
maxBudgetUsd 限制单次花费。
Related MCP server: gemini-cli-bridge
一次性 headless 执行 kscc 编码任务。
参数 | 必填 | 说明 |
prompt | ✅ | 编码任务 |
cwd | ✅ | 目标项目目录（kscc 的 cwd） |
model / effort | – | 模型别名或全名 / 努力等级 |
allowedTools / disallowedTools | – | 工具白/黑名单 |
appendSystemPrompt | – | 追加系统提示 |
permissionMode | – | 默认 |
addDirs | – | 额外可访问目录 |
outputFormat | – | 默认 |
jsonSchema | – | 强制结构化输出 |
maxBudgetUsd | – | 花费上限 |
timeoutMs | – | 默认 600000，超时杀子进程 |
noSessionPersistence | – | 不持久化会话 |
mcpConfig | – | 透传给 kscc 的 MCP 配置 |
续接会话做多轮迭代。sessionId（--resume）优先；未给则用 cwd（--continue）续接该目录最近会话。fork=true 续接时开新 session id。其余参数同上。
工具返回一个 text 块，内容为 JSON：
{"session_id":"…","text":"…","tool_uses":[],"cost_usd":0.0,"duration_ms":0,"num_turns":0,"is_error":false}session_id 用于后续 continue_kscc_session。
git clone <repo> kscc-mcp-server
cd kscc-mcp-server
npm install
npm run buildDSH 的 MCP 配置指向本 server：
{
"mcpServers": {
"kscc": {
"command": "node",
"args": ["/path/to/kscc-mcp-server/dist/index.js"]
}
}
}或免安装："command": "npx", "args": ["-y", "kscc-mcp-server"]（发布后）。
同上 stdio 配置格式。
本机已安装 kscc CLI 且在 PATH，并已完成鉴权（kscc 交互登录或 OWTFFSSENT_API_KEY）。MCP server 进程继承此环境。
Node ≥18。
npm run dev # tsx 直跑 src/index.ts
npm test # vitest
npm run build # tsc 产 dist/集成测试（test/index.test.ts）用伪 kscc 脚本，无需真实鉴权。真实 kscc 联调手动：node dist/index.js 后用 MCP 客户端调 run_kscc_prompt。
长驻 stream-json 子进程多轮复用（降延迟）。
后台 agent 管理（kscc --bg / kscc agents --json）。
HTTP 传输（多客户端 / 远程编排）。
续接 kscc 会话做多轮迭代；给 sessionId 走 --resume，否则用 cwd 走 --continue
| Name | Required | Description | Default |
|---|---|---|---|
| cwd | Yes | 项目目录；未给 sessionId 时用它续接该目录最近会话 | |
| fork | No | 续接时开新 session id，不改原会话 | |
| model | No | ||
| effort | No | ||
| prompt | Yes | 续接会话的追问内容 | |
| addDirs | No | ||
| mcpConfig | No | ||
| sessionId | No | 续接指定会话（优先于 cwd） | |
| timeoutMs | No | ||
| jsonSchema | No | ||
| allowedTools | No | ||
| maxBudgetUsd | No | ||
| outputFormat | No | ||
| permissionMode | No | ||
| disallowedTools | No | ||
| appendSystemPrompt | No | ||
| noSessionPersistence | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It reveals the internal resume/continue flag mechanics, but does not disclose side effects, budget/agent behavior, session persistence, or whether the original session is mutated (only indirectly hinted by the fork parameter in the schema). This is a significant gap for an agentic continuation tool.
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 sentence with no filler, front-loading the purpose and then giving a compact routing rule. Every segment earns its place and the structure is easy to parse.
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?
Given 17 parameters, no output schema, no annotations, and the existence of a sibling tool, this one-line description is severely incomplete. It omits return values, error behavior, side effects, cost implications, and any comparison with run_kscc_prompt, making it inadequate for reliable tool selection and invocation.
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?
Schema description coverage is only 24%, and the description only addresses sessionId and cwd—both already described in the schema. The other 15 parameters (model, effort, permissionMode, allowedTools, disallowedTools, maxBudgetUsd, etc.) receive no explanation, so the description fails to compensate for the low schema coverage.
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 action (续接/continue) and resource (kscc 会话) with a clear goal (多轮迭代/multi-round iteration). The routing clause also distinguishes this from the sibling run_kscc_prompt by making clear it resumes an existing session rather than starting a new one, even without naming the sibling.
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 gives parameter-selection guidance: with sessionId use --resume, otherwise use cwd with --continue. However, it provides no explicit guidance on when to use this tool versus run_kscc_prompt, such as 'use run_kscc_prompt for new sessions.' The usage context is implied by the word 续接 but not clearly stated.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
一次性 headless 执行 kscc 编码任务，返回结构化 JSON（含 session_id 供续接）
| Name | Required | Description | Default |
|---|---|---|---|
| cwd | Yes | 目标项目目录（kscc 的 cwd，作用域 + AGENTS.md 上下文） | |
| model | No | ||
| effort | No | ||
| prompt | Yes | 要执行的编码任务 | |
| addDirs | No | ||
| mcpConfig | No | ||
| timeoutMs | No | ||
| jsonSchema | No | ||
| allowedTools | No | ||
| maxBudgetUsd | No | ||
| outputFormat | No | ||
| permissionMode | No | ||
| disallowedTools | No | ||
| appendSystemPrompt | No | ||
| noSessionPersistence | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It transparently reveals that execution is headless, one-time, and returns structured JSON containing a session_id. This tells the agent the operation's nature and return shape, though it does not mention side effects like file modifications or cost implications. Still, it adds meaningful behavioral context beyond the schema.
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, dense sentence that front-loads the key information: one-time, headless, coding task, structured JSON, and session_id for continuation. Every part earns its place, and there is no redundant or filler content.
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?
Given the complexity of 15 parameters and no output schema or annotations, the description is far from complete. It covers the essential return behavior but does not explain the many optional parameters, when to use them, or the operational effects (e.g., file changes, budget limits). The agent would need to infer a great deal about how to invoke this tool correctly, especially for non-trivial tasks.
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?
Schema description coverage is only 13%, with only prompt and cwd having descriptions. The tool has 15 parameters, and the description provides no additional meaning for any of them. It does not compensate for the low schema coverage at all; the agent is left to guess the semantics of model, effort, addDirs, mcpConfig, permissionMode, and the rest.
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 action (one-time headless execution of kscc coding task) and a clear resource (kscc prompt). The word '一次性' (one-time) differentiates it directly from the sibling continue_kscc_session, and the mention of session_id for continuation reinforces this distinction. An agent can tell exactly what this tool does and how it differs from the sibling.
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?
Usage context is implied rather than explicit. The description notes that a session_id is returned for continuation, which hints that to continue a task you would use the sibling, but it never explicitly states when to choose this tool over continue_kscc_session or vice versa. No exclusion or alternative condition is given.
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. Dates show when Glama detected each change.
v0.1.0continue_kscc_sessionrun_kscc_promptThe two tools have clearly distinct purposes: one initiates a new coding task, the other continues an existing session. No overlap or ambiguity.
Both names follow the same 'verb_noun' pattern (run_kscc_prompt, continue_kscc_session), with consistent snake_case style.
Two tools is perfectly scoped for the server's stated purpose of executing and continuing kscc coding tasks.
The tool set covers the full lifecycle implied by the server: starting a new prompt and continuing an existing session, with no obvious missing operations.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Remote MCP server for supportsheep: run AI interviews and manage support content for your blog.
MCP server (stdio): lint OpenAPI specs with Spectral via the AgentForge API
MCP server for progressive tool usage at any scale (see https://klavis.ai)
Illumora Craft — evidence-augmented prompt compile for agents via remote MCP (OAuth) or local stdio.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/DonHz666/kscc-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server