40+ coding agent harnesses compared: loops, tools, subagents, and permissions
blaizedsouza · x · 2026-11-12
sermakarevich published part 2 of a series systematically comparing 40+ coding agent harnesses:
- How the agent loop differs across tools
- How the first request is constructed
- Toolsets and subagent mechanisms
- Output handling and permission models
Answers "what does your coding agent actually do between your prompt and the model?" and analyzes what the differences mean for users — valuable for tool selection and understanding agent mechanics.
More from coding & agent
- Alchemy author demos Z3-backed Prove/Explore for TypeScript domain invariants — samgoodwin89 · 2026-11-12
- SteerCamp user racks up 1,182 agent memory items in 2 weeks with natural recall — pramodk73 · 2026-11-12
- 3 GitHub repos worth reading for the code: nanoGPT, tinygrad, llama.cpp — cneuralnetwork · 2026-11-12
- Reddit comparison of 7 agent messaging approaches: A2A, email, Slack, custom protocols — StrongCoffee85 · 2026-11-12
- Dev feeds new Claude a Call of Duty video, ships a playable browser tactical shooter — TAbrodi · 2026-11-12
- Second V8 patch lands: TypedArray indexOf/includes up to ~30x faster on Uint8Array — cnakazawa · 2026-11-12