clew
Clew
穿越棕地迷宫的一根线。hop 就是这根线上的一个节点(docs/hops.md)。
棕地服务已经在运行。编码智能体每次会话仍要付出代价去重新发现同样的多服务流程。Clew 保留你已有的四张图——code、catalog、loadout、traces——并把它们连成一条你可以行走的路径。
Input → Branches → Output
measure → one P0 → capture → measure until reflectedgo install github.com/jeunessegamesee/clew/cmd/clew@v0.1.1无需 Scale 许可证,无需 Neo4j,无需 OpenSearch。
任意棕地,三条命令
clew init --root /path/to/repo # .clew.yaml + CLEW.md
clew stub --root /path/to/repo # IN+else for every gRPC / GraphQL / OpenAPI port
clew score --root /path/to/repo # P0 = missing IN; P1 = uncaptured branches
clew loadout --root /path/to/repo --in /package.Service/Method桩(Stub)让你可以引用每一个入站端口。没有 when 的树保持 P1 未捕获状态。用 clew loop 加深——不要凭空发明分支。
Related MCP server: Manifold Engine MCP Server
从源码构建
go test ./...
go build -o clew ./cmd/clew
./clew score --root examples/shop --fail-p0
./clew loadout --root examples/shop --use-case place-order --payload '{"paymentMethod":"CARD"}'
./clew loop --root testdata/incomplete-shop./clew score --root examples/shop --by-service
./clew score --root examples/shop --fail-p1
./clew watch --root examples/shop --traces traces/live.jsonl --follow
./clew watch --root examples/shop --listen 127.0.0.1:4318
./clew mcpclew mcp 是一个 stdio MCP 服务器(loadout、who_consumes、who_calls、next_p0、drift、score)。每个工具都有自己的输入模式。
适配文件
.clew.yaml 就是让陌生仓库变成地图的方式:
roots:
- .
- ../other-repo
services:
checkoutservice: src/checkoutservice # OTel service.name → hops folder
traces:
- traces/*.json通用入口(无需按框架解析器):OpenAPI / proto / GraphQL、OTLP(clew watch)或 clew ingest --file doors.json。可选的 http: [auto] 会猜测 Actix/Spring/……,但这不是添加新技术栈的方式。
语言
Clew 使用 Go 编写。CBM 保持为代码图。我们连接 HTTP/ASYNC/CROSS、OpenAPI/AsyncAPI 和 OTLP——我们不会重建索引器。
地图格式
参见 SPEC.md。每个服务:docs/hops.md。覆盖率:clew score --fail-p0。可执行之后:--fail-p1 会冻结未捕获的桩。
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Codebase intelligence for AI agents — dead code, blast radius, ownership.
Code intelligence platform for AI agents. 20 tools for architecture, security & impact analysis.
AI Agent with Architectural Memory. Impact analysis (free), tests and code from the graph (pro).
Codebase intelligence for agents: 152 structured artifacts across 21 programs, one call.
Related MCP Servers
- AlicenseAqualityDmaintenanceEnables AI coding agents to efficiently navigate and understand large codebases by providing tools for entry point location, call chain analysis, and impact assessment, reducing context consumption and model costs.53GPL 3.0
- AlicenseNot gradedqualityDmaintenanceEnables AI coding assistants to perform high-speed structural codebase analysis and dynamic context management by mapping repositories as byte-streams and detecting architectural chaos.MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI coding agents to analyze codebase behavior through entrypoints, call paths, and tests, providing impact analysis and context packs for code changes.41 PyPIMIT
- AlicenseNot gradedqualityDmaintenanceEnables cross-repository architecture discovery and dependency routing across multi-service codebases, with automated scanning and batch AST indexing for AI agents to trace request lifecycles and navigate service boundaries.15 npmMIT