solution-architect skill turns business problem statements and requirements into structured, modular solution architecture documentation. It bridges the gap between executive stakeholders and technical implementers by establishing clear component selection, integration patterns, and roadmaps.
Source location: skills/solution-architect/SKILL.md in the ADK repository.
Overview
When designing enterprise AI agent systems, starting with implementation details often leads to rework, misaligned architectures, or uncoordinated tool integrations. Thesolution-architect skill provides a structured methodology to analyze business needs before writing code.
Best for: Early-stage solution design, stakeholder alignment, and establishing the foundation for Standard Operating Procedure (SOP) development.
Output documents
The skill produces three focused architecture deliverables:- Executive Solution Overview — High-level summary of the business problem, target outcomes, key user personas, and ROI drivers.
- Technical Architecture & Design — System topology, agent vs. workflow decision rationale, component inventories (agents, tools, knowledge bases, integrations), and data flow specifications.
- Implementation Plan & Roadmap — Phase-by-phase implementation milestones, risk mitigation strategies, and prerequisites ready for elaboration into detailed SOPs.
Key capabilities
- Requirement Analysis: Ingests unstructured problem statements, transcripts, or RFP requirements and extracts functional and non-functional requirements.
- Architectural Decisioning: Evaluates whether capabilities belong in native agents, collaborator agents, agentic workflows, or external tools based on determinism and state requirements.
- Component Mapping: Identifies required data sources, API integrations, authentication mechanisms, and knowledge base backends (Milvus, AstraDB, Elasticsearch).
- SOP-Ready Handoff: Formats technical workflows into modular specifications optimized for consumption by the
sop-builderskill.

