1. Conduid
  2. AI
  3. LLM Agent
MCP server · AI

LLM Agent

· github.com/fr0ster/llm-agent

Core components for building Smart LLM agents, plus a default MCP-orchestrated OpenAI-compatible server implementation.

Unclaimed llmagentmcpmodel-context-protocolsap-ai-core
InstallView source
37Low

Scored 6 months ago · breakdown

About LLM Agent

LLM Agent is an MCP server in the AI category: core components for building Smart LLM agents, plus a default MCP-orchestrated OpenAI-compatible server implementation.

Install

Claude Code
claude mcp add llm-agent -- npx -y @mcp-abap-adt/llm-agent
npx
npx -y @mcp-abap-adt/llm-agent

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Ask AI

✦

Ask AI about LLM Agent

Powered by Claude · Grounded in docs

I know everything about LLM Agent. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

  • ✓README presentThe repository ships documentation.

5 checks have not been run yet.

README

Smart Agent & Server

Stand With Ukraine

A high-performance, RAG-orchestrated LLM agent and OpenAI-compatible server with deep MCP integration.

Packages

Package What it is
@mcp-abap-adt/llm-agent Core interfaces, types, MissingProviderError, lightweight helpers (CircuitBreaker, FallbackRag, LLM call strategies, ToolCache, adapters, normalizers). Zero provider dependencies.
@mcp-abap-adt/llm-agent-mcp MCPClientWrapper, McpClientAdapter, createDefaultMcpClient, and MCP connection strategies.
@mcp-abap-adt/llm-agent-rag RAG/embedder composition — makeRag (async), resolveEmbedder (sync), prefetch helpers, backend factories.
@mcp-abap-adt/llm-agent-libs Core composition runtime: SmartAgentBuilder, SmartAgent, pipeline, sessions, history, resilience, observability, plugins, skills, makeLlm/makeDefaultLlm.
@mcp-abap-adt/llm-agent-server Binary only — CLI (llm-agent, llm-agent-check, claude-via-agent) + HTTP SmartServer. Not importable as a library.
@mcp-abap-adt/openai-llm OpenAI LLM provider (OpenAIProvider).
@mcp-abap-adt/anthropic-llm Anthropic LLM provider (AnthropicProvider).
@mcp-abap-adt/deepseek-llm DeepSeek LLM provider (DeepSeekProvider, extends OpenAI-compatible).
@mcp-abap-adt/sap-aicore-llm SAP AI Core LLM provider via @sap-ai-sdk/orchestration.
@mcp-abap-adt/openai-embedder OpenAI embeddings (OpenAiEmbedder).
@mcp-abap-adt/ollama-embedder Ollama embeddings + RAG (OllamaEmbedder, OllamaRag).
@mcp-abap-adt/sap-aicore-embedder SAP AI Core embeddings (SapAiCoreEmbedder).
@mcp-abap-adt/qdrant-rag Qdrant vector store RAG (QdrantRag, QdrantRagProvider).
@mcp-abap-adt/hana-vector-rag SAP HANA Cloud Vector Engine RAG (HanaVectorRag, HanaVectorRagProvider). Optional peer.
@mcp-abap-adt/pg-vector-rag PostgreSQL + pgvector RAG (PgVectorRag, PgVectorRagProvider). Optional peer.

Quick install

(a) Server-managed declarative (most common)

Install server + exactly the peers your smart-server.yaml references:

# Fully local — Ollama LLM + Ollama embeddings
npm install @mcp-abap-adt/llm-agent-server \
            @mcp-abap-adt/ollama-embedder

# DeepSeek LLM + Ollama embeddings
npm install @mcp-abap-adt/llm-agent-server \
            @mcp-abap-adt/deepseek-llm \
            @mcp-abap-adt/ollama-embedder

# SAP AI Core LLM + SAP AI Core embeddings + Qdrant RAG
npm install @mcp-abap-adt/llm-agent-server \
            @mcp-abap-adt/sap-aicore-llm \
            @mcp-abap-adt/sap-aicore-embedder \
            @mcp-abap-adt/qdrant-rag

A missing peer throws MissingProviderError at startup with an install hint.

(b) Programmatic composition (no YAML)

Install llm-agent-libs + the peers you need. Construct SmartAgent via SmartAgentBuilder in code. Import provider classes from their packages and pass instances to the builder's fluent setters.

npm install @mcp-abap-adt/llm-agent-libs \
            @mcp-abap-adt/llm-agent-mcp \
            @mcp-abap-adt/llm-agent-rag \
            @mcp-abap-adt/deepseek-llm \
            @mcp-abap-adt/ollama-embedder

(c) Core-only (no SmartAgent, no server)

npm install @mcp-abap-adt/llm-agent

Build your own agent against the interfaces exported by core. Supply your own ILlm and IEmbedder implementations.

See docs/MIGRATION-v11.md if you are upgrading from v10.

Documentation

  • QUICK_START.md — end-to-end guide: install, config, connect IDE
  • ARCHITECTURE.md — architecture reference: thin proxy layer + SmartAgent/SmartServer/pipeline
  • INTEGRATION.md — custom interface implementation guide with code examples
  • PERFORMANCE.md — RAG, BM25, model selection, token budget tuning
  • CLIENT_SETUP.md — connection instructions for Claude CLI, Cline, and Goose
  • SAP_AI_CORE.md — SAP AI Core operational guidance and troubleshooting
  • DEPLOYMENT.md — production deployment patterns (Docker, systemd, serverless)

Development

# Build project
npm run build

# Run tests
npm run test:all

# Development with hot-reload
npm run dev

# Smart server production entrypoint
npm run start

# Legacy compatibility aliases
npm run start:smart
npm run dev:llm
npm run start:llm
npm run test
npm run test:llm

License

MIT

README copy captured when this server was indexed; it may lag the repository. The original is authoritative.

Questions

About LLM Agent

How do I install LLM Agent?

Run claude mcp add llm-agent -- npx -y @mcp-abap-adt/llm-agent, then add the server to your MCP client's configuration.

Is LLM Agent safe to use with an AI agent?

Its trust score is 37 out of 100 (low). It passes 1 of the 1 static security checks run so far. It has no ConduID identity yet, so agent calls to it are not receipted.

Is LLM Agent still maintained?

Conduid hasn't recorded a commit date for this repository yet. Check the repository directly for recent activity.