MCP ServerSTDIOOfficialv1.10.11

Keshrath Agent Tasks MCP Server

Manage multi-stage task pipelines for AI agents with dependency tracking, artifact handling, and task claiming. Built for developers orchestrating complex workflows where agents need to coordinate work across sequential or parallel stages.

io.github.keshrath/agent-tasks

Hosted URL

Local install

Transport

STDIO

Auth

No auth required

What the Keshrath Agent Tasks MCP server does

How models use it and what it is built for.

Manage multi-stage task pipelines for AI agents with dependency tracking, artifact handling, and task claiming. Built for developers orchestrating complex workflows where agents need to coordinate work across sequential or parallel stages.

Connect to Keshrath Agent Tasks

Local install — runs as a subprocess.

npx agent-tasks@1.10.11

Resources

Where to find authoritative docs and source for Keshrath Agent Tasks.

Example prompts for Keshrath Agent Tasks

Paste any of these into Agent Studio after connecting Keshrath Agent Tasks.

  • Create a pipeline with three stages: data_fetch, process, and validate with dependencies between them
  • Claim an available task in the queue and retrieve its input artifacts
  • List all pending tasks and their current stage status across the pipeline
  • Mark a task complete and output the result artifacts for downstream stages

Keshrath Agent Tasks MCP server — FAQ

Common questions about connecting and running Keshrath Agent Tasks.

  • What is agent-tasks used for?

    agent-tasks is a task management system designed for AI agents to orchestrate multi-stage pipelines. It handles stage definitions, task dependencies, artifact passing between stages, and task claiming so multiple agents can coordinate work without conflicts.

  • How do I install and run agent-tasks?

    Install via npm with `npx agent-tasks@1.10.11`. It runs as a stdio-based MCP server, meaning it communicates through standard input/output streams with your AI agent or client application.

  • Can I run tasks in parallel or only sequentially?

    agent-tasks supports both: you define dependencies between stages, so tasks can run in parallel when they have no blocking dependencies, or sequentially when one stage must complete before the next begins.

  • How does task claiming work?

    Agents can claim available tasks from the queue to prevent duplicate work. Once claimed, a task is locked to that agent until it completes or is released, ensuring only one agent processes it at a time.

  • What happens to artifacts between pipeline stages?

    Artifacts are passed between stages as task output. When a task completes, its artifacts become available as input to dependent downstream tasks, enabling data flow through the pipeline.

Skip the local setup — run MCP in your browser

MCP Playground runs 10,000+ hosted MCP servers — GitHub, Linear, Notion, Stripe, Sentry and more — across Claude, GPT, Gemini, DeepSeek and 30+ AI models. Compare model answers side-by-side, save agent presets, share runs. Zero install.

Open Agent Studio

Related servers

More on MCP Playground