Skip to main content
Glama

智能体系统

一个持久化的多智能体编排器,具有以下特性:

  • 显式运行图和检查点/恢复功能

  • 编排器控制的并行委派

  • 有界研究集群执行

  • 编码、审查、修复、CI 和审批循环

  • 通过 MCP 资源和工具暴露的后端存储中的项目内存

  • 用于内存检索的真实 SQLite 向量索引

  • 支持 Tavily 的可插拔外部研究后端

范围

此实现针对 MCP 2025-11-25 规范基准,使用官方 Python MCP SDK 和用于内存表面的 FastMCP 服务器。对于本地开发,它通过 stdio 运行。对于远程部署,请参阅 docs/remote_auth.md。

Related MCP server: ellmos-homebase-mcp

布局

  • app/runtime: 运行状态、调度器、编排器循环、检查点

  • app/planner: 规划和图修订辅助工具

  • app/agents: 用于研究、代码、审查、修复、CI、综合、审批的节点执行器

  • app/memory: 基于 SQLite 的内存、检索和工件索引

  • app/mcp_server: FastMCP 资源、工具、提示词和服务器入口点

  • tests: 验收和单元测试覆盖

本地使用

uv sync --group dev
uv run pytest
uv run agent-system-mcp

检索和研究后端

  • 内存条目使用 sqlite-vec 索引到本地 SQLite 向量表中。

  • 默认嵌入提供程序为 auto:它优先使用真实的 sentence-transformers 模型,仅在模型无法加载时回退到确定性哈希提供程序。

  • 当节点提供 inputs.corpus 时,研究使用内存语料库后端。

  • 如果设置了 TAVILY_API_KEY,无语料库的研究节点可以使用 Tavily 后端进行外部网络研究。

  • 如果既没有语料库也没有 Tavily 密钥,研究将返回有界的空结果,而不是编造来源。

嵌入配置

  • AGENT_SYSTEM_EMBEDDING_PROVIDER=auto|sentence-transformers|hash

  • AGENT_SYSTEM_EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2

  • AGENT_SYSTEM_EMBEDDING_CACHE_DIR=/path/to/cache

  • AGENT_SYSTEM_EMBEDDING_LOCAL_ONLY=true|false

示例:

AGENT_SYSTEM_EMBEDDING_PROVIDER=sentence-transformers uv run agent-system create-run "improve scheduler"

本地传输

开发使用 MCP stdio 传输。

远程部署

远程部署在 v1 版本中仅提供文档说明。服务器记录了兼容 OAuth 2.1 的同意路径,并保留本地 stdio 作为默认开发模式。

完整文档

Available Tools

10 tools
checkpoint_runB
Idempotent

Apply a small state patch and create an explicit checkpoint.

Inputs: run_id (required), state_patch_json (required). Returns: object

create_runA

Create a new orchestrator run from a high-level goal and optional JSON spec.

Inputs: goal (required), spec_json. Returns: object

get_file_contextB
Read-onlyIdempotent

Read file context from the orchestrator workspace.

Inputs: paths (required), include_neighbors. Returns: object

plan_goalB
Read-onlyIdempotent

Generate a large-app execution blueprint from a high-level goal and optional JSON spec.

Inputs: goal (required), spec_json. Returns: object

record_decisionC
Idempotent

Record a high-value architectural or implementation decision.

Inputs: title (required), run_id (required), node_id (required), rationale (required), +2 more. Returns: object

record_failure_modeA
Idempotent

Record a reusable failure pattern or bug signature.

Inputs: run_id (required), node_id (required), summary (required), signature (required), +1 more. Returns: object

resume_runA
Idempotent

Resume an orchestrator run until completion, pause, or failure.

Inputs: run_id (required). Returns: object

search_project_memoryA
Read-onlyIdempotent

Search project memory by query string and optional type filters.

Inputs: query (required), kind, limit, status. Returns: result

write_project_memoryC
Idempotent

Write a durable memory entry with provenance to the backing store.

Inputs: key (required), kind (required), run_id (required), node_id (required), +5 more. Returns: object

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 10 tools

    Disambiguation3/5

    Most tools target distinct actions or resources, but plan_goal and create_run both take a high-level goal and optional JSON spec, making their boundaries unclear. Similarly, search_project_memory and list_related_artifacts both retrieve stored information in overlapping ways. Descriptions help clarify intent, but some pairs could still lead to misselection.

    Naming Consistency5/5

    All tools follow a consistent verb_noun snake_case pattern, such as resume_run, write_project_memory, record_decision, and create_run. The naming is uniform and predictable across the entire toolset.

    Tool Count5/5

    Ten tools is a well-scoped count for an orchestrator and memory management server. Each tool covers a meaningful operation without the set feeling bloated or overly sparse.

    Completeness4/5

    The core lifecycle is covered: creating and resuming runs, checkpointing, planning, memory read/write, and recording decisions/failures. Minor gaps exist, such as no explicit run status/query tool or delete operation for memory entries, but agents can generally work within the provided surface.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      D
      maintenance
      Shared memory and orchestration for coding agents, enabling persistent knowledge, multi-agent coordination, and a canonical workflow across MCP-compatible AI clients.
      83
      -
    • A
      license
      C
      quality
      B
      maintenance
      Enables local-first LLM orchestration with persistent memory, knowledge management, routing, swarm patterns, API probing, tests, automation planning, and plugin discovery via a stdio MCP server, using SQLite for offline storage.
      51
      173 npm
      2
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      A production-grade multi-agent workflow orchestrator built on the Model Context Protocol, featuring planner/executor/critic agents, durable run state, and replayable traces.
      15
      MIT
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables autonomous multi-agent orchestration using directed acyclic graph workflows to interact with external developer infrastructure such as Git, Postgres, and Docker through safety-gated MCP servers.
      Apache 2.0