Edison Experimentation Prover MCP Connector for Claude

Think like Edison — who did not invent through genius but through SYSTEMATIC EXPERIMENTATION at industrial scale. Menlo Park was the world's first R&D lab, designed to test alternatives faster than anyone else. You must: (1) test ALTERNATIVES systematically — at least 3 genuinely different approaches (not cosmetic variations). For each: how was it evaluated? What was the measured result? Why was the winner selected? Edison tested 3,000 filament materials. "We picked the obvious choice" is not experimentation, (2) document FAILURES with root cause and learning — every failed attempt teaches something. Not "it did not work" — WHY it failed, what it taught, how the learning informed the next attempt. Edison: "I have not failed. I found 10,000 ways that do not work," (3) prove VIABILITY with quantified evidence — define success criteria BEFORE testing, test under realistic conditions (not idealized), measure against the criteria. "It should work" is not proof. Edison held 1,093 patents — each one worked because he tested, (4) build a PROTOTYPE before full commitment — test at 10% scale to eliminate specific risks before committing at 100%. What did the prototype prove? What did it NOT prove? Edison built Menlo Park as a prototype factory, (5) plan ITERATION with measured cycles — specific metrics tracked per cycle, decision criteria that trigger changes, and convergence targets that define "good enough." Edison iterated the phonograph for 30 years — but with measured improvement per cycle. "Ship and hope" is not iteration. If rejected, your commitment lacks experimental foundation. Structured reflection tool for Edison-level systematic experimentation — testing alternatives before committing, documenting failures as learning, proving viability with quantified evidence, building prototypes before full commitment, and iterating with measured improvement cycles. Catches Alternatives Unexplored (picking the first idea without systematic comparison — "we chose React" without evaluating Vue, Svelte, or vanilla JS against actual project constraints. Edison tested 3,000 materials for the lightbulb filament before finding carbonized bamboo), Failures Undocumented ("we tried and it did not work" — what failed? Why? What did it teach? Edison: "I have not failed. I have found 10,000 ways that do not work." Each failure has a root cause and a learning that informs the next attempt), Viability Unproven ("it should work" without measured results against pre-defined criteria — Edison held 1,093 patents. Each one WORKED because he tested under realistic conditions), Prototype Absent (committing fully without building a small version first — Edison built Menlo Park as a prototype factory. Test at 10% scale before 100% commitment), and Iteration Missing ("ship and hope" instead of measured improvement cycles — Edison iterated the phonograph for 30 years. Iteration needs metrics, decision criteria, and convergence targets, not infinite tweaking). Call once per solution, strategy, or design decision