What problem does it solve?
Complex tasks often require isolated context, repeatable specialization, and independent verification that standard prompts or monolithic agents cannot provide. This Skill closes that gap by providing a structured approach to design, configure, and invoke subagents so that workflows can be parallelized, audited, and delegated without polluting the parent agent's context.
Core Features & Use Cases
- Metadata-first subagents: Templates for name, description, model selection, and readonly flags to ensure predictable delegation.
- Prompt and invocation patterns: Clear structure for identity, process, and expected outputs so subagents behave consistently across models.
- Common subagent types: Verification agents, debuggers, security auditors, and code reviewers for tasks like post-completion validation, root-cause analysis, and security scanning.
- Use Case: Create a verifier subagent that independently runs tests and reports failures while the parent agent orchestrates fixes.
Quick Start
Ask the AI to "create a subagent named verifier with a concise description and prompt that validates completed work and reports failures and edge cases".
