What problem does it solve? Writing a well-structured agent prompt from scratch is slow and error-prone. This Skill gives you a curated library of 18 tested prompt templates so you can adopt a proven agent definition instead of designing one yourself. ## Core Features & Use Cases - Three template categories: general-purpose agents (code-reviewer, support-assistant, data-analyst, technical-writer, security-auditor, api-designer), task agents (summarizer, translator, extractor, classifier, generator, validator), and specialized-domain agents (legal, medical, financial, marketing, HR, educational). - Standard agent format: every template uses the Kimi Code YAML frontmatter (name, description, model, color, tools, skills) plus a structured Markdown body with output formats and explicit Forbidden rules. - List, search, view, and customize: browse templates, inspect one before adopting it, and copy it with customizations such as language or focus. - Use Case: You need a PR review agent. Instead of writing one, run the library's code-reviewer template, which already encodes SOLID/OWASP checklists, severity-categorized feedback (BLOCKING/SUGGESTION/NITPICK), and a structured verdict format. ## Quick Start Ask the assistant to list the prompt library templates and use the code-reviewer template for your next pull request review.
