DeepSeek Open-Sources DSH: A Fully Pluggable Agent Framework for Self-Evolution
量子位 · wechat · 2026-08-14
DeepSeek has officially open-sourced its native Agent framework, DeepSeek Harness (DSH). The framework adopts an "everything is a plugin" architecture, modularizing models, tools, strategies, and context management, allowing developers to customize runtimes like Lego blocks.
DSH features multiple built-in modes (Standard, PTC, Minimal, Creative) and natively supports event-level trajectory replay, Token consumption, and cache hit rate monitoring. Tests show strong resilience and autonomous iteration in long-horizon tasks like 3D modeling and game dev. Furthermore, DSH introduces the Cordis protocol, enabling models to write and install plugins autonomously during tasks, aiming to explore AI "self-evolution" via a community plugin ecosystem.
Related event: DeepSeek Open Sources Plugin-Based Agent Framework Harness v0.1(36 posts)→
More from coding & agent
- edith-1 monitors agent traces, beats Sonnet-5.5 on balanced accuracy at 1/274 the cost — xennygrimmato_ · 2026-10-09
- How edith-1 works: probabilistic filter plus expensive agent judge for flagged runs — xennygrimmato_ · 2026-10-09
- Splash 1.3.0 cuts local agent first-token latency from 19s to 1s via SSD offloading on M6 Mac — BeidiChen · 2026-10-09
- OpenAI makes coding-agent steering instant, ships new "GPT-6.1 Sol ultrafast" release — Dimillian · 2026-10-09
- Robotics dev: LLM users keep spawning new robot libraries instead of improving existing ones — kscottz · 2026-10-09
- Modal Launches Sidecars: Sandboxed Containers for Agent-Generated Code with 3x Faster Cross-Trust-Boundary Communication — AAAzzam · 2026-10-09