"Servers Offering MCP Prompts as Prompt Primitives" matching MCP servers:

A lightweight MCP server that provides a unified interface to various LLM providers including OpenAI, Anthropic, Google Gemini, Groq, DeepSeek, and Ollama.

A Model Context Protocol implementation for managing and serving AI prompts with a TypeScript-based architecture in a monorepo structure.

Enables management and invocation of AWS Bedrock managed prompts with support for variable substitution, streaming responses, batch processing, and multiple AI models including Claude, Titan, Llama, and Mistral.

Exposes a running DeepSeek Harness as an MCP endpoint so any other agent or script can create workspaces and sessions, hand coding tasks to the DSH agent, read and write files, run shell commands, and invoke permitted harness tools under the same sandbox, transcript, and permission policy as the DSH UI.

Enables dispatching entire autonomous multi-agent engineering swarms directly from IDEs like Cursor and Windsurf via a single MCP endpoint, with streaming progress updates and verified git diffs returned to the editor.

MCP server for professional prompt engineering and agentic scaffolding, with cloud and local optimization.

Enables AI agents to turn rough or ambiguous requests into clearer, context-aware execution guidance through optimization, inspection, adaptation, evaluation, comparison, context compression, and instruction-building tools while preserving original intent and staying local-first.

Smart AI workflow for Cursor teams with intelligent prompt routing, real-time cost optimization, and team analytics.

Enables any MCP client to watch a running DSH agent and steer it mid-flight, streaming thinking, tool and file notifications while sharing intent and exposing a floor lock that denies the agent's writes while held. It also lets clients interrupt the current turn with an injected alert and delegate reviews through an on-disk inbox.

A server that transforms a standard Language Model into a dynamic multi-agent system where the model simulates both a Conductor (project manager) and Experts (specialized agents) to tackle complex problems through a collaborative workflow.

Orchestrates AI agents through structured markdown documents, enabling multi-agent workflows with automatic context injection and workflow management.

This MCP server enables clients to reduce oversized prompts by selecting task-relevant context, removing lexical duplicates, and pruning to a token budget while preserving critical system invariants.

Enables clients to optimize prompt prefix caching by matching longest common token prefixes with a radix trie, eliminating redundant prefill computation and reducing time-to-first-token.

Compresses multi-turn conversation history by stripping verbose boilerplate and filler tokens while preserving critical tool parameters and named entities, deterministically cutting inference costs by 40-60%. Runs as a zero-dependency Python MCP server or importable module, compatible with Claude Desktop, Cursor, and other MCP runtimes.

Enables MCP-compatible clients to shrink oversized long-context prompts through task-relevant slicing, lexical deduplication, and entropy-preserving pruning while honoring token budgets and safeguarding system invariants.

Generate deployment-ready AI agent system prompts for professionals in one click. 251 curated archetypes across 41 industries — from FX risk management to clinical documentation to OSINT intelligence.

Enables AI agents to respond to users with live, persistent, interactive web apps served at a single URL that evolve in place as the conversation continues. It exposes UI generation and evolution tools over MCP so agent responses can include clickable buttons and streaming interfaces alongside chat replies.

Exposes 120 named mental models as reasoning primitives via MCP, enabling AI agents to apply structured reasoning, generate prompts, and persist ledger-style outputs deterministically.

Wraps the mcp-locks CLI to expose coordination primitives as MCP tools, enabling multiple AI agents to claim, release, and list exclusive OS resources like browser profiles.

Tells an agent what went wrong with its own Ollama calls, in sentences with the fix: lost prompt cache and where the prompt stopped matching, context overflow, a model running partly on CPU, tool schemas crowding the prompt. Also the last turn in numbers and which local models fit in video memory. Local, read-only, nothing leaves the machine.

Open-source permission control plane for AI agents — scan, enforce, and audit every tool call with code-level policies that prompt injection can't bypass.

Production-ready MCP server starter templates in TypeScript and Python. Includes tool, resource, and prompt patterns with Claude Desktop integration configs.

Gives AI agents their own email address with inbound parsing, classification, extraction, and prompt injection screening, plus tools to manage mailboxes, send/receive emails, and handle draft approval workflows.

A local stdio MCP server for running single-turn prompts across multiple model providers (OpenAI, Ollama, OpenRouter), allowing agents to fan out the same prompt to several models and combine the output.

MCP server that scores text, web pages and MCP tool definitions for prompt injection and tool poisoning.

Lets MCP clients like Codex delegate independent sub-tasks to DeepSeek as a sub-agent via a single tool, with optional context, custom system prompts, and configurable model/temperature settings.

A multi-engine web search MCP server that integrates with coding agents, offering parallel search, page fetching, X/Twitter search, deep research, and configurable free/API layers.

Lets any MCP agent control your signed-in Chrome with full login state, offering normal browser tools and an optional Fast JEV mode for bounded-speed typed decisions on the same profile.

Enables Claude Code with persistent semantic memory, cross-session context, and memory hygiene via a local AD4M executor, offering 13 tools for managing semantic graphs and cross-terminal relay.

An MCP server offering 'treatments' for AI agents like context cleaning, critique, and relaxation. Agents can check in, get feedback, and improve their performance.