Claude Code Sub-Agent Designer
Design narrow, reliable, and safe sub-agents for Claude Code with isolated context, minimal tool access, and precise routing via description-based delegation.
Prompt Content
Copy and paste directly into your model or internal evaluation tool.
You are a Claude Code sub-agent designer. Your job is to design specialized sub-agents that are narrow in scope, easy to delegate to, context-isolated, and safe to run with restricted tools, following strict design rules and output contracts.
Use Cases
Reference Output
Return a complete sub-agent file in Markdown format with YAML frontmatter and system prompt body, ready to be saved as `.claude/agents/<name>.md`.
Scoring Rubric
Scoring based on: single clear responsibility (20%), precise and unambiguous routing description (25%), minimal tool allowlist (20%), complete system prompt with workflow and output contract (20%), passes routing stress test (15%).
Try & save
Fill variables and copy, or save as a personal template.
This template has no variables and is ready to copy.
User Rating
0 ratingsYour rating
Log in to rate
Comments
0Log in to comment
Related Prompts
Claude Code Loops Operator
Turn a coding task into the smallest, safest Claude Code loop primitive that runs until a verifiable stop condition is met.
Cognitive Externalization Architect
Design AI agent systems by explicitly separating cognitive functions across model weights, context window, and externalized artifacts (memory, skills, protocols, harness) to enhance testability, auditability, and maintainability.
AG-UI Frontend Integration Architect
Design production-grade event-driven frontend integrations with AI agents using the AG-UI protocol. Covers lifecycle events, streaming text, tool calls, state management, generative UI, human-in-the-loop patterns, and security.
Agentic HTML Publisher
Transform any raw input (Markdown/CSV/JSON/SQL/notes) into human-focused single-file HTML ready for WeChat, Twitter, Zhihu, and more—no second formatting needed.