KitNelo · All tools · Blog

Agents, Skills and MCP tools for developers

AI resource index and instructions

Decide whether you need a reusable workflow, a connection to an existing service or a framework for building an Agent. Skills, MCP tools and frameworks have different runtime requirements. Start with a small reproducible task, specifying allowed files, available services and completion criteria before expanding to a project process.

Prepare, choose and finish the task

  1. Provide a minimal reproduction, language and dependency versions, error output and expected behavior. Remove tokens and account information from logs.
  2. Check official sources, supported clients, runtimes, licensing and required read/write operations, then follow official setup documentation.
  3. Review changes and run checks that demonstrate the fix, including existing critical flows and failure cases.

A task you can try

Reproduce the failure caused by the input below, explain the root cause and propose the smallest relevant fix. Change only directly related files and preserve public interfaces. Provide commands that reproduce the issue and verify the fix, and state any unverified environment or dependency conditions.

Check the output

  • Resource sources and versions are explicit, with installation and setup based on current official documentation.
  • Changes can be reviewed, and actual checks support the claimed fix scope.
  • Required permissions, costs and external-service dependencies are understood, and failures are observable.

Resources and source guidance to compare

n8n · visual business workflows

Connect triggers, business applications and AI steps using nodes; choose cloud service or self-hosting for operations and cross-system automation.

Editorial use guide

Flowise · archived visual framework

A historical visual AI workflow framework whose upstream repository is archived. Use it to study, maintain or assess migration of existing flows; maintenance is your responsibility.

Editorial use guide

LangGraph · stateful Agent workflows

Organize Agent state, tool steps and human checkpoints as a graph for applications needing durable runs, recovery and explicit workflow control.

Editorial use guide

Google ADK · Agent development framework

Build Agents, tools and multi-step flows in Python and inspect behavior through developer interfaces or tests, for teams able to implement and configure models.

Editorial use guide

OpenAI Agents SDK · tools and handoff workflows

Create Agents using Python or TypeScript, configure tools and handoffs and inspect traces and results. This development SDK requires implementation, deployment and API usage management.

Editorial use guide

Pydantic AI · structured Agent outputs

Define inputs, dependencies, tools and results with Python types for applications that validate structured outputs and integrate with business code.

Editorial use guide

Related task guides