AI Prompts & Agents Library

This library consists of some of the most commonly used usecases, select a usecase from the left pane and a corresponding optimal LLM prompt definition, an Agent definition example, a SKILL.md file template and a Plugin template will be generated, which you can tweak further if needed to customize them to your needs.

  1. Prompt : A well structured, detailed prompt for your usecase is generated that you copy & paste it into your LLM of choice.
  2. Agent Definition : An example agent code (using LangGraph) that you can use to implement for your use-case.
  3. SKILL.md file example : An example folder structure and SKILL.md template is generated for the use-case.
  4. Plugin : An example claude plugin folder is created for the use-case.

For more details regarding each of the above 4 constructs, why and where they are used - refer to the Applied AI Section under AI Engineering tab.

1. Prompt definition for this use case

This section shows a Prompt example for the selected use case using a reliable prompt engineering framework (described in the next section)

No use case selected yet

Prompt Engineering Framework used

Every prompt in this library is built using a simple, repeatable framework. For each use case we clarify the model's role, the context it sees, the task, the expected output structure and any guardrails.

  1. Role assignment: Select a use case on the left to see how we define the model's role for that prompt.
  2. Context injection: Select a use case on the left to see what information the model expects from you.
  3. Task clarity: Select a use case on the left to see a concise description of the task for that prompt.
  4. Output structure: Select a use case on the left to see the expected response format.
  5. Guardrails and constraints: Select a use case on the left to see the safety and reliability constraints we apply.

Framework adapted from Parloa: Prompt engineering frameworks.

2. AI Agent definition for this use case - LangGraph example

This section shows a simple LangGraph-style agent definition for the selected use case. It mirrors the same system prompt and intent as the text prompt above and is meant as a starting point for a multi-agent system.

No use case selected yet

Agent template building framework

When I convert a prompt into an agent, I like to keep these building blocks in mind:

  1. Goal Interpretation: The ability to understand user intent and translate it into a specific, measurable objective.
  2. Planning: Breaking down complex goals into sequential or parallel steps with built-in fallbacks.
  3. Tool Use: The capability to interact with external systems (APIs, databases, web search) to take real-world action.
  4. Observation & Iteration: A feedback loop where the agent assesses the outcome of its actions and adapts its next step.

3. SKILL.md file example for this use case

Skills are a simple, open format for giving agents new capabilities and expertise. Agent Skills are folders of instructions, scripts, and resources that agents can discover and use to do things more accurately and efficiently.

No use case selected yet

References:
https://agentskills.io/home
https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices

4. Plugin files examples for this use case

Plugins let you extend Claude Code with reusable skills, agents, hooks, and MCP servers—custom functionality you can share across projects and teams.

No use case selected yet

References:
https://github.com/anthropics/claude-plugins-official
https://code.claude.com/docs/en/plugins