What problem does it solve? When improving one aspect of a system worsens another (e.g., faster startup vs. higher memory use), the first idea that fits is often a compromise. This Skill forces three genuinely different solution candidates onto the table before planning, so trade-offs are resolved deliberately instead of by accident. ## Core Features & Use Cases - Contradiction-Driven Divergence: States the contradiction in technical or physical form, checks the Ideal Final Result, and inventories reusable code before proposing anything new. - Deterministic Principle Suggestion: The triz_suggest.py script keyword-ranks ~20 software-mapped TRIZ principles, picks one per separation mode (time, space, condition, level), and falls back to a generic trio when nothing matches. - Knowledge Graph Integration: Surfaces prior decisions that resolved similar tensions and captures new resolutions tagged with contradiction, separation mode, and principle. - Use Case: A nav-brief flags "improving rollback safety vs worsening codebase cleanliness" on a substantial task; the Skill produces three candidates (temporary shims, sidecar module, opt-in flag), each with cost and downside, then recommends one. ## Quick Start Ask the assistant to find alternatives for a contradiction, for example: "triz this: improving test coverage vs worsening CI runtime".