AI harnesses explained in 6 terms: agent loop, tool calling, context, checkpoints, sandboxing, guardrails
femke_plantinga · x · 2026-10-13
Femke Plantinga breaks down what an AI harness actually does, using a support agent that checks an order and drafts a reply:
- Agent loop: model picks an action, a tool runs, result returns; cycle ends on completion, help needed, or limits
- Tool calling: model requests tools by name/args; harness validates, executes, returns results
- Context management: prompts, hooks and compaction decide what the model sees each step
- Checkpointing: workflow state saved outside the model so interruptions are resumable
- Permissions & sandboxing: scope tools, files and code execution (read orders, no refunds)
- Guardrails: input/output/action checks and approval gates — none catch every error
Payoff: swap models across labs while sharing the same skills and context team-wide.
More from coding & agent
- Daniel Jeffries: only ~2% can actually use AI agents, mass adoption is 10-20 years away — Dan_Jeffries1 · 2026-10-13
- Fully local AI short film mimics Wes Anderson style with a detailed open pipeline — Mist3r-Robot · 2026-10-13
- Controlling his Mac's terminal from a phone via t3 code beta while waiting at speech therapy — NERDDISCO · 2026-10-13
- LLM cache verifier drift: reject-only calibration tripled error rate; 5% audits fix it — Reasonable_Royal_621 · 2026-10-13
- A style guide for making Claude-like models write normally, and its linguistic traps — gleech · 2026-10-13
- Open-source desktop agent dodaios runs local Qwen models via ollama — and shows where 16GB models fail — InternationalGap3698 · 2026-10-13