Building production agents: system prompts over role play, reasoning models by default
PawelHuryn · x · 2026-09-09
Steps 1-2 of the author's production-agent thread.
- Step 1 — define a system prompt: it sets goals, autonomy boundaries, and ways of working. Tip: skip the role, explain why the work matters and the strategic context, and progressively disclose more via INDEX.md, skills, etc. Links free guides including Anthropic's prompt engineering docs.
- Step 2 — pick the model: start with a reasoning model unless the framework handles iterating (e.g., n8n). The author says GPT-5.6 Luna is by far the most cost-efficient in his benchmarks—since costs scale with users, default to it.
Related event: Building Production Agents: System Prompts First, Workflows as Tools(2 posts)→
More from coding & agent
- xbrlkit: open-source XBRL layer over Arelle with built-in MCP server support — jfrench009 · 2026-09-09
- Dev Adds a Complete Tech Tree With GPT-6 Astra in 2h21m — Dimillian · 2026-09-09
- Cheap models via OpenRouter fall apart in agentic harnesses: GLM and DeepSeek can't match Claude — scottyLogJobs · 2026-09-09
- Multi-agent coding's hardest problem: deciding who is allowed to change what — apghere · 2026-09-09
- Tip: have your coding agent maintain a cross-linked markdown wiki under docs/ for persistent decisions — chris_j_paxton · 2026-09-09
- Keep a markdown wiki under docs/ so coding agents remember decisions across sessions — chris_j_paxton · 2026-09-09