emircbngl
UserClaude Code skill that runs procurement-grade research on any product — requirements, fair-price discovery, TCO, risk, decision memo, lifecycle plan. Region-neutral, B2C + B2B, self-feeding domain knowledge base.
Categories
Indexed Skills (2)
procurement
Research any consumer or business product (bicycles, PCs, cameras, audio gear, smart-home, appliances, cosmetics, furniture, mobility, IT software, medical devices, and anything else — unknown categories auto-derive a new domain pack and save it for future runs) as a professional procurement specialist would: formalize requirements, build a candidate slate, validate standards compliance and compatibility, run TCO and vendor-risk analysis, and deliver a documented decision memo with a post-purchase lifecycle plan. Self-feeds — builds up its own domain-pack library as it researches new categories. Supports both B2C (personal / household) and B2B (company / procurement) modes. Auto-extracts context from working-directory files and asks the user only for what's missing. Triggers on: "research [product]", "compare [A] vs [B]", "help me pick a [category]", "is X compatible with Y", "what should I upgrade on my [thing]", "should I buy [product]", "build me a [PC/bike/setup] for [budget]", "evaluate [product] for [us
optiproof
Optimize a function and PROVE the speedup — like code review, but for performance. YOU (Claude Code) write candidate implementations; the `optiproof` harness proves each is behaviourally identical (differential testing) and statistically faster (Mann–Whitney + bootstrap CI), then keeps the verified-fastest. No API key needed — the agent is the candidate generator, optiproof is the prover. Triggers on: "optimize this function", "make this faster", "speed up <func>", "/optiproof", "optimize <file>::<func> and prove it".
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.