ost AI agent libraries are markdown files in a folder. Operator Agents is different. Every agent ships with a structured YAML schema defining: • tools.required — what the agent needs to run • context.required — what inputs you must provide • output.format — json, markdown, code, or mixed • pipeline.handoff_to — which agent runs next + condition This makes agents composable. An orchestrator can read agent-index.json and route tasks automatically.

Krisp Accent Conversion — Understand accented speech in real time

Understand accented speech in real time

Hey PH! 👋 I'm Fatih, automation architect and the builder behind Operator Agents. Quick story: I kept building multi-agent automations and hitting the same wall — agent libraries gave me prompts but no structure. No way to know what an agent needs, what it outputs, or what should run next. So I designed AGENT_SPEC.md — a YAML schema every agent must follow. Then built 66 agents following this spec. The key insight: if agents declare their tools, context requirements, output format, and pipeline handoffs in machine-readable YAML, you can build orchestrators that work on the schema, not just the prompt text. Would love your feedback, especially: - The pipeline handoff protocol (is the JSON contract approach the right abstraction?) - Which verticals are most useful to you - Runtime support gaps Happy to answer anything — I'll be here all day! 🚀 GitHub: https://github.com/fatihkutlar/o...

Krisp Accent Conversion — Understand accented speech in real time

Understand accented speech in real time

Operator Agents

66 AI agents with machine-readable YAML schemas

66 AI agents with machine-readable YAML schemas