Agent Architect
Use when converting a product idea, workflow, or rough requirement into a concrete AI agent architecture with purpose, scope, loop, tools, memory, permissions, and success criteria.
Permission Reviewer
Use when reviewing an agent's permissions, allowed tools, dangerous operations, approval points, secret handling, or security boundaries.
Multi Agent Orchestrator
Use when splitting a complex task across multiple agents or designing parent-agent/subagent collaboration, reviewer agents, worker agents, and handoff protocols.
Handoff Writer
Use when creating a concise continuation note for another agent, another developer, or a future session after partial completion of a long-running task.
Tool Contract Designer
Use when designing, reviewing, or documenting tools that an AI agent can call. Produces typed tool contracts with schemas, permissions, failure modes, and audit rules.
Verification Loop Designer
Use when defining how an agent proves that its work is correct, including tests, checks, evidence, acceptance criteria, and failure recovery.
Context Pipeline Designer
Use when designing what an agent should load, retrieve, summarize, ignore, or persist across a task or long-running workflow.
Agent Harness
Use when designing, reviewing, or improving an AI agent, coding agent, research agent, workflow agent, support agent, or multi-agent system. Applies when the user asks for agent architecture, tool design, context management, permissions, verification, handoff, orchestration, or an installable/open-source agent skill package.