Prototype-to-codebase feasibility analysis via Claude and a read-only MCP knowledge base
OK_darein · reddit · 2026-11-11
A Reddit user proposes an AI-assisted feature feasibility workflow: feed an interactive HTML prototype (e.g. generated with Claude) to Claude alongside extracted technical knowledge from existing repos and API docs, exposed via a custom read-only MCP server, to produce a report on reusable components, required frontend/backend changes, API support, risks, and rough effort.
The stated goal is understanding technical implications before development—not generating code. The main hurdle is that PMs often lack access to repositories or Swagger/OpenAPI specs, hence the knowledge-base layer. The author is soliciting existing implementations, courses, and MCP/RAG lessons rather than presenting a finished solution.
More from coding & agent
- Workflow: get AI to design multiple Dynamic Island options first, then write native code — PrajwalTomar_ · 2026-11-11
- OpenAI Releases 34-Page Whitepaper on Building, Evaluating and Deploying AI Agents — mdancho84 · 2026-11-11
- One-file playable stealth heist game built by Claude, all assets drawn in code — heypearlai · 2026-11-11
- Building an AI data analyst: how to handle conflicting per-user KPI definitions — Lopsided_Feedback_81 · 2026-11-11
- 7 memory slots vs 200K tokens: the hardware mismatch behind AI coding debt — esrtweet · 2026-11-11
- Sydney Codex Community Buildathon set for Nov 14 with Cerebras-backed ultrafast compute track — paw_lean · 2026-11-11