Atlas / MCP servers / mybolide / Probe Kit

Probe KitBLOCK

Name: Probe Kit
Author: mybolide

mcp/mybolide/probe-kit

MCP server by mybolide · ai · ai agent · ai tools · Audited 2026-10-08

Spec-driven dev workflow MCP for Cursor & AI agents — start_feature/bugfix orchestration, auto Skill+AGENTS bootstrap, memory, GitNexus. npx -y mcp-probe-kit@latest

Verdict

BLOCK

Grade

Trust score

69 /100

Exposed tools

64 50r · 13w · 1d

Transport

stdio

License

MIT

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

知时MCP | mcp-probe-kit Know the Context, Feed the Moment.

Introspection · Context Hydration · Delegated Orchestration

Talk is cheap, show me the Context. mcp-probe-kit is a protocol-level toolkit designed for developers who want AI to understand project intent, choose a precise workflow, and retain validated experience without flooding the model with internal actions.

[](https://www.npmjs.com/package/mcp-probe-kit) [](https://www.npmjs.com/package/mcp-probe-kit) [](https://opensource.org/licenses/MIT) [](https://github.com/mybolide/mcp-probe-kit/stargazers) [](https://mseep.ai/app/ce79a9e5-7422-4862-b235-bc3d65710397)

🚀 AI-Powered Complete Development Toolkit - Covering the Entire Development Lifecycle

A powerful MCP (Model Context Protocol) server with 29 model-visible tools by default (24 base + 5 History Session; disable History with MCP_HISTORY_SESSION=0), 35 when Memory is also configured, and a 39-tool compatibility surface available through MCP_TOOLSET=full. It covers the complete workflow from product analysis to final release and supports structured ou

Read from source at commit d16a86f114adOBSERVED · 2026-10-08

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code (npm)

claude mcp add mcp-probe-kit -- npx -y [email protected]

Exposed tools (64)

50 read · 13 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`AGENTS.md`	read	项目 Agent 规则与 mcp-probe 上下文块
`AppError`	read	统一错误封装
`add_feature`	write	原子规格生成工具:根据已经确定的 feature_name、description、spec_layout 和可选 subspecs,返回 flat 或 parent-child 规格模板与落盘计划。它不负责判断需求复杂度,也不是复杂功能的首个入口。新功能、跨模块升级或布局未确定时必须先调用 start_feature;通常仅按 start_feature 返回的 delegated plan 调用本工具。
`agent-auth`	write	Add account authentication
`architecture`	read	独立架构领域能力,使用 ARC-8 完成架构评估、设计、校验和漂移检查。可直接调用,也可由功能、Bug 或重构流程按需组合;MCP 负责方法、门禁与结构化证据,不替 Agent 声称绝对最优架构。
`ask_user`	read	当 AI 需要更多信息、遇到不确定因素时使用。向用户提问工具,AI 可主动向用户提问;支持单个或多个问题、提供选项、标注重要性;可在任何时候使用
`audit-start-feature`	write	Add account authentication
`blue-saas`	read	常见 B2B 蓝色主色,清晰可信
`check_spec`	read	校验已落盘的功能规格(docs/specs/<feature_name>/requirements\|design\|tasks.md)是否完整:检测残留 [填写] 占位、缺失章节、缺 FR/验收标准、FR 未进覆盖矩阵。写完规格后、进入实现前调用;未通过按报告补全后重跑。
`code_insight`	read	当用户需要基于代码图谱分析调用链、上下文和影响面时使用。默认桥接 GitNexus,支持 query/context/impact 模式;不可用时自动降级并返回原因
`code_review`	read	当用户需要审查代码、真实 Git diff 或托管 Plan 的交付一致性时使用。MCP 可确定性收集 changed files、Plan 声明范围、产物、测试、公共契约、架构和 revision 证据;代码语义问题仍由 Agent 审查,不伪装成静态扫描器
`commerce-v2`	read	v2 兼容升级
`commerce-v3`	read	商业系统版本升级
`converge`	read	按 Delegated Plan 自己声明的证据和质量闸门关闭计划。任一未完成步骤、未决事项、必需证据或验收结果缺失都会拒绝收敛;除 requirements 外,证据只有摘要但没有 reference/revision 也不算可复核证据。调用参数只能增加证据要求,不能削弱 Plan。
`delete_memory_asset`	destructive	按 asset_id 从共享记忆库删除一条未参与 supersede 链的资产。适用于错误、重复或无价值沉淀;已有关联关系的资产必须改用 update_memory_asset 设置 retracted 并保留 evidence。
`demo-auth`	read	用户认证:支持邮箱登录、注册、密码重置;需要会话保持与失败锁定
`estimate`	read	当用户需要估算开发工作量、评估任务时间时使用。估算开发工作量,输出故事点、时间范围(乐观/正常/悲观)、风险点
`fix_bug`	read	当用户需要找问题、修 bug、排查异常时使用。指南型 SRC-8(TBP-inspired):注入真因工作表与门禁,由 Agent 完成 rootCauseAnalysis 并修复;MCP 不自动修 Bug
`full-agent-audit`	read	Compatibility specification fixture
`full-audit-feature`	read	Compatibility spec fixture
`gencommit`	write	当用户需要生成 Git commit 消息时使用。返回 Conventional Commits 规范说明、步骤、输出模板和示例,供 AI 根据变更内容生成最终 commit message。它不直接代写最终消息,也不应被判定为空结果
`gentest`	read	当用户需要为代码生成单元测试时使用。指南型工具:注入 code/file_path 与测试清单,由 Agent 生成完整测试代码;MCP 不自动生成或运行测试
`green-health`	read	医疗/健康类产品的绿色主色
`history_session_bootstrap`	read
`history_session_checkpoint`	read	Append an idempotent, redacted development checkpoint. Pass session_key and expected_path exactly as returned by history_session_bootstrap, plus the user\
`history_session_read`	read	Read one lossless numeric Markdown archive by number or a path returned from history_session_search. Responses are UTF-8-safe pages: max_bytes defaults to 32 KiB and is capped at 64 KiB; follow next_cursor to recover the complete source.
`history_session_search`	read	Search lossless history archives by deterministic keywords and return a bounded page of ranked locations and snippets. Use history_session_read to retrieve exact source text.
`history_session_validate`	read	Validate history-session numbering, duplicates, and derived state/manifest. Optional repair rebuilds state and manifest from Markdown archives.
`init_project`	read	当用户提供一句话需求时使用。基于 Spec-Driven Development 理念,分析需求并生成完整的项目规格文档(需求分析/技术设计/任务拆解)。适合项目初期的需求澄清和规划
`init_project_context`	read	生成/更新项目上下文写作计划(delegated):MCP 写入 AGENTS.md 与 layout.json;project-context 分类文档与 graph-insights 由 Agent 按返回的 plan 落盘。新功能请先 start_feature,修 bug 请先 start_bugfix。
`interview`	read	当用户需求不明确、需要澄清需求时使用。需求访谈工具,在开发前通过结构化提问澄清需求,避免理解偏差和返工;生成访谈记录文件供后续 start_feature/add_feature 使用;仅支持 feature 类型
`list_memory_assets`	read	Memory Center 专用浏览动作。按更新时间列出记忆摘要,支持类型、状态、项目、标签和分页过滤;仅供 MCP App 调用。
`mcp-v4`	read	跨模块协议、工具注册、Memory、Task Runtime 和兼容性矩阵的多阶段升级
`memory-degradation-stability`	read	开发一个简单功能并验证 Memory 服务不可用时仍能生成 delegated plan
`memory-feature`	write	Add login with reusable validation
`memory-smoke-asset`	read	本地 smoke test 资产
`not_registered`	read
`orange-energy`	read	活动/增长类产品的橙色主色
`plan_heartbeat`	read	由 Agent 在 Delegated Plan 执行过程中记录轻量检查点。首次调用必须提供完整 plan;后续合并步骤、证据、产物、候选经验、验收结果、运行证据和 revision。除 requirements 外,每条用于收敛的 evidence 必须至少提供 reference 或 revision,否则 converge 会明确拒绝。只记录状态,不代替 Agent 执行。
`read_memory_asset`	read	当编排阶段已检索到记忆摘要,且 AI 需要查看完整沉淀代码或详细规范时使用。根据 asset_id 读取记忆资产详情。
`refactor`	read	当用户需要重构代码、改善代码结构时使用。指南型工具:注入 code/file_path 与重构清单,由 Agent 分析后输出重构计划 JSON;MCP 不自动修改源文件
`resume_plan`	read	从 .mcp-probe-kit/plans/ 读取 Delegated Plan 检查点,按依赖计算下一可执行步骤、阻塞步骤和 resumeContext。plan_id 可选;省略时自动恢复当前项目最近更新的 active/blocked Plan。本工具只读取状态,不替 Agent 执行;返回 mustContinue=true 后 Agent 必须立即执行 nextStep/nextTool,逐步调用 plan_heartbeat,禁止只汇报恢复结果后停止。
`rose-beauty`	read	美妆/生活方式产品的柔和玫瑰色
`scan_and_extract_patterns`	read	当需要从单段代码、单文件或整个目录中抽取可复用模式,再决定是否沉淀到记忆系统时使用。目录扫描时,优先传 project_root 为项目根目录绝对路径,并让 directory_path 传相对项目根的路径,例如 app/utils;只有无法确定项目根时,才把 directory_path 直接设为绝对路径。不要传带项目名的半相对路径,例如 font-miniapp-api/app/utils。${SCAN_PATTERN_PERSIST_HINT}
`search_memory`	read
`slate-dashboard`	read	偏冷灰的数据看板与分析产品
`start_bugfix`	write	当用户需要找问题、修 bug、排查异常时使用。默认按 SRC-8(TBP-inspired)编排:收敛边界→真因工作表→修复→测试→记忆沉淀。
`start_feature`	write	新功能、功能增强、大版本升级或跨模块研发的首选入口。Agent 必须把当前对话已确认的完整目标、范围、模块、阶段和约束汇总到 description;用户只说“继续/开始/往下做”时不得原样透传。默认 spec_layout=auto,复杂多模块或多阶段需求会先生成 parent-child 子规格拆分计划,再进入 add_feature→check_spec→实现。仅在规格布局和子规格已明确、且只需渲染规格模板时才直接用 add_feature。
`start_onboard`	write	当用户需要快速上手新项目时使用。编排:生成上下文文档。
`start_product`	write	产品设计完整工作流入口。返回闭环 delegated plan:Agent 生成 PRD 与原型文档,调用 ui_design_system 和 start_ui 完成设计系统及可交互 HTML 原型,并更新项目上下文;不会引用不存在的 gen_prd/gen_prototype 工具。
`start_ralph`	write	用于需要多轮小步实现、每轮真实验证和正式收敛的长任务。返回有界 Delegated Plan、每轮 Heartbeat 证据契约和可选前台辅助脚本;不自动运行循环、不创建后台进程。安全停止不等于成功
`start_ui`	write	编排 UI 设计与实现:先锁定视觉方向和信息架构,再生成关键页面,后续通过真实截图评分与迭代完成验收。
`stone-editorial`	read	内容/媒体站点常用的暖灰 editorial 感
`sync_ui_data`	write	同步 UI/UX 数据到本地缓存。来源:uipro-cli、shadcn/ui registry、内嵌 UI 主题预设、Vercel Web Interface Guidelines。
`text-only-crud`	read	Host text-only CRUD verification asset
`ui_design_system`	read	生成可执行的视觉方向,而不是风格标签拼盘。输出核心任务、信息架构、内容密度、排版与色彩策略、组件原则、明确禁用项和截图验收标准。
`ui_search`	read	搜索页面结构、组件、交互规范和实现参考。新 UI 流程优先使用 structure 模式按任务和页面类型选择信息架构;旧 search/catalog/template 模式继续兼容。
`update_memory_asset`	write	按 asset_id 更新已有记忆资产(保留原 ID)。可修正内容、证据、适用边界、失效时间及替代关系;content 或生命周期语义变化会重新向量化。${MEMORY_SHARED_KNOWLEDGE_RULE}
`user-auth`	read	用户认证功能
`violet-creative`	read	偏创意工具/设计产品的紫色主色
`workflow`	read
`zinc-neutral`	read	默认中性 SaaS 风格,低饱和、专业克制
`服务器状态`	read	MCP Probe Kit 服务器当前状态
`项目上下文索引`	read	AGENTS.md 或 docs/project-context.md 入口

Trust audit

BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

Layer	What it checks	Result
L0	Provenance & inventory	FAIL
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (7 observation(s))
Network: none-observed
Shell: declared (5 observation(s))
Dependencies: not all pinned
Secrets in source: none-found

Findings (23)

CRITICALObfuscation / stealth · obf.decode_then_exec · CWE-506, CWE-94

src/resources/generated/mcp-app-bundle.ts:2

atob( ... new Function(

Why it matters. decodes a payload and executes it

HIGHInventory / provenance · inv.suspicious_name · CWE-1104

src/lib/memory-payload.ts

memory-payload.ts

Why it matters. member named after an attack tool

Fix. remove or justify

MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59

delete_memory_asset

Why it matters. 1 tool(s) can delete or overwrite

Fix. prefer a read-only mode or scoped tokens; the page states the blast radius

LOWInventory / provenance · inv.suspicious_name · CWE-1104

src/lib/__tests__/memory-payload.unit.test.ts

memory-payload.unit.test.ts

Why it matters. member named after an attack tool

Fix. remove or justify

LOWCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95

src/tools/__tests__/code_review.unit.test.ts:93

fs.writeFileSync(filePath, "eval(input);\n", "utf-8");

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table

LOWCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95

src/tools/__tests__/code_review.unit.test.ts:98

expect(text).toContain("eval(input)");

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table

LOWInsecure crypto · crypto.weak_hash · CWE-327, CWE-338

src/tools/__tests__/start_product.unit.test.ts:23

product_name: 'rc4 acceptance product',

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

scripts/lib/tool-surface-baseline.mjs:3

const DEFAULT_BASELINE_URL = new URL('../../config/tool-surface-baseline.json', import.meta.url);

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/cli/__tests__/command-line.unit.test.ts:6

import { VERSION } from "../../version.js";

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/lib/__tests__/cli-fallback-installer.unit.test.ts:6

import { VERSION } from "../../version.js";

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/lib/__tests__/handles.unit.test.ts:6

} from '../../lib/handles.js';

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/lib/__tests__/harness-adapters.unit.test.ts:13

import { VERSION } from "../../version.js";