Did you know you can completely rewrite OpenCode’s default system prompt directly from your config? All you have to do is drop a prompt key into your opencode.json file and point it to a local markdown file. You can target the build agent, the plan agent, or use this exact architecture to spin up entirely custom sub-agents inside your repo. Here is what the configuration looks like: from: Agents | OpenCode { "$schema": "https://opencode.ai/config.json", "agent": { "build": { "mode": "primary", "model": "anthropic/claude-sonnet-4-20250514", "prompt": "{file:./prompts/build.txt}", "permission": { "edit": "allow", "bash": "allow" } }, "plan": { "mode": "primary", "model": "anthropic/claude-haiku-4-20250514", "permission": { "edit": "deny", "bash": "deny" } }, "code-reviewer": { "description": "Reviews code for best practices and potential issues", "mode": "subagent", "model": "anthropic/claude-sonnet-4-20250514", "prompt": "You are a code reviewer. Focus on security, performance, and maintainability.", "permission": { "edit": "deny" } } } }

Hi there! I'm new in this community. I am preparing myself to pass the CCA-f (Claude Certified Architecte), anyone has any tips or ressources to be ready to success it ?

I was just looking at how many members we have in here and we are so close to hitting 400!!😲 I wanted to say a quick thank you to everyone who has joined! It really means a lot to me that you have decided to be a part of this journey with us and allow us to be a part of your journey as well! Me and @William Guidry are planning a lot of things for the community and we are excited for this next chapter!!

SuperPowers + PonyTail + Graphify = amazing agent skill workflow combo superpowers: obra/superpowers: An agentic skills framework & software development methodology that works. PonyTail: DietrichGebert/ponytail: Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote. Graphify:safishamsi/graphify: AI coding assistant skill (Claude Code, Codex, OpenCode, Cursor, Gemini CLI, and more). Turn any folder of code, SQL schemas, R scripts, shell scripts, docs, papers, images, or videos into a queryable knowledge graph. App code + database schema + infrastructure in one graph.

A company spent $1.2 million on an AI platform. Six months later, the frontline employees built a workaround in Excel. A spreadsheet beat a million-dollar platform. And it wasn't even close. I keep seeing the same fight play out. One side says "change the culture first, then bring the tools." The other side says "just ship the tools, the culture will follow." Both sides have PowerPoints. Neither side is listening. Here's what actually determines which one comes first: you have to figure out whether your people are resisting the CHANGE or resisting the TOOL. Those are completely different problems. A manufacturing company I worked with rolled out an AI system that could predict equipment failures 72 hours in advance. Technically brilliant. The maintenance crew ignored it completely. Why? Nobody told them it was coming. Nobody asked what problems they actually had. And the first time the system flagged a false positive, the crew chief said, "See? Told you. Computers don't know machines." Done. Whole team wrote it off in one sentence. A year later, same company brought in a change management consultant. Six months of workshops, ride-alongs, small pilots. Then they reintroduced the exact same tool. Same platform. Same algorithms. Adoption hit 78%. Same tool. Different culture. Completely different result. But before you think culture-first always wins, here's the other side. A former client of mine spent eight months on "AI readiness." Surveys. Town halls. Change champions. Eight months of preparation. Zero AI in production. Competitor advantage: gone. Meanwhile, another firm in the same industry just picked one team, handed them an AI tool, and said "try it for two weeks." No fanfare. No transformation workshop. No forty-seven-slide deck with a stock photo of a handshake. Two weeks later, that team asked for more licenses. So here's the diagnostic I use with clients. 3 questions: Previous tech rollouts failed? If yes, culture first. You're fighting scar tissue from the last time leadership promised a miracle tool.