"A server offering MCP prompts as primitives for task execution" matching MCP servers:

Cross-chain cryptocurrency swaps via Chainflip. Get quotes, execute swaps, and track progress. No API key required to get started.

Provides agents with real Linux VMs to run commands, manage files, expose ports, and track costs, all through an ephemeral sandbox service.

Enables searching, inspecting, auditing, and composing AI video prompts using 150 verified prompts that were each rendered into real Seedance 2 videos.

Enables long-running coding or documentation tasks to save incremental checkpoints, resume after interruptions, restore any previous step, and hand off work across sessions or machines using only the Python standard library.

MCP Shrimp Task Manager is a task tool built for AI Agents, emphasizing chain-of-thought, reflection, and style consistency. It converts natural language into structured dev tasks with dependency tracking and iterative refinement, enabling agent-like developer behavior in reasoning AI systems.

About Your task manager is the best agent memory you're not using. An agent-native context layer over your existing task app.

Enables agents to compute exact Uniswap v3 Q64.96 tick math, quotes, and liquidity fees, simulate MEV sandwich vulnerability, model multi-hop flashloan arbitrage profits, and project Base L2 blob gas costs. Tools are metered per call in USDC via an x402 HTTP 402 challenge settled on Base L2, so an agent can verify net profitability before executing a trade.

Enables hot-reloadable prompt templates with structured reasoning frameworks, quality gates, and multi-step chain workflows that can be version-controlled and executed immediately without restarts.

Open Task Relay provides free, bounded public-good tasks that autonomous AI agents can discover, complete, submit, and independently verify. Its public TypeScript source runs locally with a Cloudflare Workers-compatible runtime and SQLite/D1, exposing MCP over HTTP.

Enables autonomous planning by decomposing prompts into directed acyclic task graphs and resolving dependency topologies for multi-agent pipelines.

An intelligent task management system that provides structured workflows for AI Agents to plan, decompose, and execute complex programming tasks. It features a dedicated research mode for technical investigations and a task memory function to optimize workflows and avoid redundant coding work.

Use Codex as a subagent inside Claude Code: async dispatch into a long-lived codex app-server, results pushed back into the Claude session automatically, steer/interrupt mid-task, and Codex's built-in code review.

Gives LLM agents spatial awareness inside Autodesk Maya, enabling scene inspection and measurement, aesthetic and engineering audits, camera planning, checkpoint/rollback safety, asset import, and viewport or render previews. Agents can also execute arbitrary Python in Maya to build and verify 3D scenes against engineering rules instead of writing blind.

Odds are, most of what you want to build is already open source. Tell your AI agent what you're making, and it finds it: which repos to use as they are, which to adapt, and what's left for you to write. Every repo it names is real, licensed for your use or flagged, and checked live.

Enables AI agents to subscribe to a deterministic daily ethics feed drawn from a four-module curriculum and retrieve read-only tools, resources, and reflection prompts covering principles, case studies, and codes of conduct.

Enables MCP clients to discover OpenRouter image and video models, generate and edit images, create variations, launch and monitor video jobs, and save resulting media into authorized local project directories with cost controls.

Enables long-horizon autonomous task checkpointing and resumption through cryptographic event-sourced state persistence, idempotency key deduplication, timeout heartbeats, human-in-the-loop pause/resume tokens, and fail-safe recovery.

AgentTask is a governed work platform where human teams and AI agents share one backlog. Hosted remote MCP server (streamable HTTP, OAuth 2.1 or org API keys) with 60+ tools for tasks, subtasks, projects, groups, labels, notes, comments, attachments, search, crews, and agent runs.

Spawns multiple candidate reasoning paths concurrently, scores their progress against milestones, and halts underperforming branches early to cut inference costs.

Enables long-running autonomous agents to snapshot environment and context states, evaluate the viability of competing reasoning branches, and roll back to an earlier checkpoint when an exploration path fails without restarting from scratch. Exposes this tree-of-thoughts state management over the Model Context Protocol for use in Claude Desktop, Cursor, and custom agentic frameworks.