45 agent skills from shipping 3 products, plus cost-per-score math on Gemini 4 Neon
PawelHuryn · x · 2026-11-20
Product engineer Pawel Huryn open-sourced his product-engineering repo (MIT, 661 stars): 45 agent skills distilled from shipping three live products — decide what to build, build with Claude Code or Codex, have a second model review, release.
His personal setup: Opus 5.5 (xhigh) for trust, strategy, taste, and managing other agents; GPT-6.2 Sol for coding and reviews. Ships /delegation, /codex-agents, /gemini-agents and more.
He also did score-vs-cost math: Gemini 4 Neon would be ideal at 63/105 for $1.80 per run at xhigh (57 for $1.13 at high) — but Google only permits it via API key; running it from other harnesses violates the terms (allegedly being changed).
More from coding & agent
- Bug Hunt chart: score vs subscription cost, Gemini 4 Neon xhigh tops at 63 for $1.80 — PawelHuryn · 2026-11-20
- Measured: Claude Max delivers 46.4x API value on Opus 5.5, beating ChatGPT Pro's 10.25x — PawelHuryn · 2026-11-20
- 10 open-source GitHub repos that turn your docs into queryable knowledge graphs — victor_explore · 2026-11-20
- What happens when two AI agents disagree? The multi-agent conflict problem — EricJose59 · 2026-11-20
- AI turns Mediterranean illustration into interactive 3D world; Three.js wall-crash demo costs $6.98 — Saboo_Shubham_ · 2026-11-20
- Higgsfield launches GPU Harness: rent GPUs inside Claude and ChatGPT to render and train LoRAs — CurieuxExplorer · 2026-11-20