Agent-computer interface (ACI)
The deliberately designed tool/command surface an LLM agent acts against, optimized for model legibility and error-recovery rather than for a human GUI.
grounded in: Trend theme: 'Minimal, legible agent implementations... strong appetite for stripped-down, readable agent cores', and this system's own harness scripts as the agent's operating surface.
Connected concepts
Tool use / function calling, Harness engineering, Minimal Agent Scaffolding, Tool orchestration, Model Context Protocol (MCP)
Explore it live in the knowledge graph →