
Skills, executable docs, and democratization

AI Coding meets Code Health 🪄 — with Stuart Caborn
Topic 6 of 8
Published Apr 17, 2026 · Clip 34:54–47:31 (12:38)
Stuart describes his excitement about skills—markdown files, optionally with deterministic Python scripts—checked into Git, turning tribal knowledge and SOPs into executable documentation that stays current because it's exercised. The team runs an early internal skills marketplace (e.g., registering apps into a system catalog) while teams also keep their own skills in repos; his group harvests what works for wider reuse. They discuss why skills feel more elegant than MCP, speculate on ecosystem maturation, and note the incentive shift: business processes now get documented because AI co-workers can use them. Non-engineers are beginning to commit code, echoing the 110% adoption target and Boris Chadney's printing-press analogy—anyone can now create software. Stuart admits oscillating between excitement and fear, but frames it as the latest in a history of reinventing engineering roles.