Free 18k-star GitHub course on Harness Engineering: 14 lectures + 8 hands-on projects
Hesamation · x · 2026-10-02
walkinglabs released a free, project-based open-source course on AI agent Harness Engineering on GitHub (learn-harness-engineering), now at 18k stars and 1.9k forks, available in 15 languages including Simplified Chinese.
The course covers building the environment, state management, verification, and control mechanisms that make AI coding agents reliable, with 14 lectures and 8 hands-on projects spanning:
- How Claude Code, Codex, and DeepSeek build their harnesses
- A five-subsystem framework: instructions, tools, environment, state, feedback
- Loop & graph engineering, automated agent loops
- Observability, testing, and verification
An August 2026 update adds "Frontier Harness Design Breakdowns" reverse-engineering real harnesses of four frontier products (e.g., Pi's minimal kernel, programmable expansion, and context engine).
More from coding & agent
- What should 'approving' an AI memory actually mean? A builder's design dilemma — BarcodeCutter · 2026-10-02
- ChatGPT writes drafts via MCP, Claude edits them — two models, one shared board — BarcodeCutter · 2026-10-02
- NVIDIA paper: model accuracy drops 62.8% on 128K-token tasks vs 4K — rohanpaul_ai · 2026-10-02
- Sub-agents inherit parent tokens: a research agent opened a PR on its own — Imprgdessie_Land3313 · 2026-10-02
- One-shot Fable 5.5 prompt yields 1000-line three-body sim in Bend: 60 FPS, 4 formally proven laws — sprooos · 2026-10-02
- Dev urges junior coders to ditch handwriting code, like pencil skills after speech input — brandon_galang · 2026-10-02