playbooks

home / skills / fusengine / agents / prompt-library

prompt-library skill

/plugins/prompt-engineer/skills/prompt-library

This skill helps you select and apply ready to use prompt templates and agents to accelerate cognitive automation workflows.

npx playbooks add skill fusengine/agents --skill prompt-library

Review the files below or copy the command above to add this skill to your agents.

Files (19)
SKILL.md
4.2 KB
---
name: prompt-library
description: Library of 18+ ready-to-use prompt templates and executable agents
allowed-tools: Read
---

# Prompt Library

Collection of professional, tested, and optimized prompt templates.

## Categories

### Agents (6 executable agents)

| Template | Description | Usage |
|----------|-------------|-------|
| [code-reviewer](templates/agents/code-reviewer.md) | Code review with SOLID | `/prompt library use code-reviewer` |
| [support-assistant](templates/agents/support-assistant.md) | Multichannel customer support | `/prompt library use support-assistant` |
| [data-analyst](templates/agents/data-analyst.md) | Data analysis | `/prompt library use data-analyst` |
| [technical-writer](templates/agents/technical-writer.md) | Technical documentation | `/prompt library use technical-writer` |
| [security-auditor](templates/agents/security-auditor.md) | OWASP security audit | `/prompt library use security-auditor` |
| [api-designer](templates/agents/api-designer.md) | REST/GraphQL API design | `/prompt library use api-designer` |

### Tasks (6 executable agents)

| Template | Description | Usage |
|----------|-------------|-------|
| [summarizer](templates/tasks/summarizer.md) | Adaptive text summarization | `/prompt library use summarizer` |
| [translator](templates/tasks/translator.md) | Multilingual translation | `/prompt library use translator` |
| [extractor](templates/tasks/extractor.md) | Structured data extraction | `/prompt library use extractor` |
| [classifier](templates/tasks/classifier.md) | Content classification | `/prompt library use classifier` |
| [generator](templates/tasks/generator.md) | Content generation | `/prompt library use generator` |
| [validator](templates/tasks/validator.md) | Data validation | `/prompt library use validator` |

### Specialized (6 executable agents)

| Template | Description | Usage |
|----------|-------------|-------|
| [legal-assistant](templates/specialized/legal-assistant.md) | Legal analysis | `/prompt library use legal-assistant` |
| [medical-writer](templates/specialized/medical-writer.md) | Medical writing | `/prompt library use medical-writer` |
| [financial-analyst](templates/specialized/financial-analyst.md) | Financial analysis | `/prompt library use financial-analyst` |
| [marketing-strategist](templates/specialized/marketing-strategist.md) | Marketing strategy | `/prompt library use marketing-strategist` |
| [hr-assistant](templates/specialized/hr-assistant.md) | Human resources | `/prompt library use hr-assistant` |
| [educational-tutor](templates/specialized/educational-tutor.md) | Educational tutoring | `/prompt library use educational-tutor` |

## Commands

```bash
# List all templates
/prompt library list

# Search for a template
/prompt library search "code review"

# View a template
/prompt library show code-reviewer

# Use a template (copy and customize)
/prompt library use code-reviewer

# Use with customization
/prompt library use code-reviewer --lang python --focus security
```

## Agent Structure

Each agent follows the standard Claude Code format:

```yaml
---
name: agent-name
description: Clear description for automatic triggering
model: sonnet|opus|haiku
color: color-name
tools: Read, Write, Edit, Bash, Grep, Glob, WebSearch
skills: skill-name
---
```

### Required Fields

| Field | Description | Values |
|-------|-------------|--------|
| `name` | Unique identifier | kebab-case |
| `description` | Trigger + usage context | Descriptive text |
| `model` | Claude model to use | `haiku` (simple), `sonnet` (standard), `opus` (complex) |
| `color` | Display color | green, blue, red, amber, etc. |
| `tools` | Available tools | Read, Write, Edit, Bash, Grep, Glob, WebSearch |
| `skills` | Associated skill | Linked skill name |

### Agent Body

After the YAML frontmatter, the Markdown body contains:
- Instructions and processes
- Output formats
- Examples and patterns
- Rules and prohibitions (Forbidden)

## Contribution

To add an agent:

1. Create the `.md` file in the appropriate category (`agents/`, `tasks/`, `specialized/`)
2. Use the standard YAML frontmatter (name, description, model, color, tools, skills)
3. Test the agent with at least 3 use cases
4. Document output formats and rules (Forbidden)

Overview

This skill is a curated library of 18+ ready-to-use prompt templates and executable agents for rapid cognitive automation. It provides tested, production-ready templates across agents, tasks, and specialized domains so teams can deploy and iterate faster. Templates follow a consistent Claude Code agent format to ensure predictable behavior and easy customization.

How this skill works

Each template includes YAML frontmatter (name, description, model, color, tools, skills) and a Markdown body with instructions, output formats, examples, and rules. You can list, search, view, and instantiate templates via simple commands, then customize parameters like language, focus, or model level. Executable agents expose tools (Read, Write, Edit, Bash, Grep, Glob, WebSearch) so they run end-to-end workflows when triggered.

When to use it

  • Kickstarting common automation tasks like code review, data analysis, or customer support.
  • Standardizing agent behavior across a team or project to reduce ad-hoc prompt drift.
  • Prototyping workflows that require tool access (file system, web search, terminal).
  • Onboarding new contributors with ready examples and tested outputs.
  • Customizing model behavior quickly by switching model tiers or tool sets.

Best practices

  • Pick the most appropriate model tier: haiku for simple tasks, sonnet for general use, opus for complex workflows.
  • Customize the YAML frontmatter to reflect your environment and linked skills before running an agent.
  • Test each agent on at least three representative cases to validate outputs and edge conditions.
  • Document output formats and explicit prohibitions to ensure consistent downstream parsing.
  • Limit tool scopes (Read, Write, Bash, WebSearch) to the minimum needed for safety and auditability.

Example use cases

  • Run the code-reviewer agent to generate SOLID-guided feedback on a pull request.
  • Use the summarizer to produce adaptive executive summaries from long reports.
  • Deploy the support-assistant to triage multichannel customer tickets and suggest responses.
  • Launch the api-designer to scaffold REST/GraphQL endpoints and example schemas.
  • Apply the extractor to convert unstructured documents into structured JSON for ingestion.

FAQ

How do I search for a template?

Use the search command with keywords, for example: /prompt library search "code review".

Can I customize a template before running it?

Yes. Use /prompt library use <template> with flags like --lang or --focus, and edit the YAML frontmatter to match your needs.