prompt-library_skill

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.