Writing architecture skill files for agents beats any prompt trick for cold starts
Successful-Tax1306 · reddit · 2026-10-20
The author shares a team practice: every new coding-agent session used to waste time poking around the codebase and guessing wrong about how services fit together. The fix was writing short skill files describing the architecture — which service owns what, how data flows between them, and which patterns are current vs. legacy code being migrated away from.
Now anyone can point a design or coding agent at those files and it starts from the right place instead of guessing. The upfront cost pays back on every new task. The author asks the community: one big context file, many small per-area files, or letting the agent explore each time — and how do you keep them from going stale?
More from coding & agent
- Ethan Mollick's new book charts 30 months of AI progress, from cyborg to conjurer work — rseroter · 2026-10-23
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Jensen Huang details Vera Rubin agentic stack: NVL72, MGX, CMX and RTX — rohanpaul_ai · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23
- Capy 0.5.0 lets Claude Pro/Max subscribers reuse their plan quota, adds OIDC cloud access — garrytan · 2026-10-23