aetox-skills/senior-architect-agent ? reverse-engineered prompt

Build me a reusable AI architecture skill that helps an agent inspect an existing codebase before suggesting changes.

I want it to map the system first, then produce clear architecture notes, module and data flow summaries, Mermaid diagrams, a debt and risk review, open questions, and a handoff note for another agent or engineer. It should always start with the smallest safe pass, ask questions when the scope is unclear, and be honest about what it knows versus what it is inferring.

Please include the docs, rules, templates, examples, and any small scripts or adapter files needed so this feels like a complete usable skill package. Make the output good for real world existing systems, not just toy examples, and make sure it can handle both a quick scan and a deeper full system review when needed. If you need to check current docs online while setting it up, go ahead.