把 OpenSpec(规范驱动开发)接入 DeepSeek Harness:内置六个 agent skill,并从插件自身依赖解析 openspec CLI,无需全局安装、无需按项目拷贝 .agents/skills | OpenSpec (spec-driven development) for DeepSeek Harness: six bundled agent skills + CLI resolution from the plugin dependency. No global install.
Stop pasting API keys into DeepSeek Harness — reuse the OAuth tokens your local Codex CLI and Claude Code already hold as model routes, with subscription usage built in.
Model pricing & capability board for DeepSeek Harness: ~7,250 models across 213 providers, honest lowest-listed comparison, session costs priced by the route you actually called, and cache-leak attribution. Local-first, no server.
A manual Terminal tab for the DeepSeek Harness (dsh) web UI — run commands on the host machine, persistent cwd, sudo password prompt, command history.现在可以在web界面内直接执行命令行了
Minimal, verified template for DeepSeek Harness plugins: bundle manifest, one tool, runtime peer guard, tests, and CI that really invokes the tool (dsh 0.1.0-rc.6).
Multi-folder workspaces for DeepSeek Harness: one workspace spanning several folders, with sessions able to read and write every member. Core patch set + plugin + a one-command npx launcher.
Reusable Markdown prompt profiles for DeepSeek Harness with per-turn model selection, argument substitution, and state restoration. DeepSeek Harness 可复用 Markdown Prompt Profile,支持单轮模型选择、参数替换和状态恢复。
DeepSeek Harness agent preset for adversarial review: presumes the artifact is wrong, demands a citable authority and file:line for every finding, refuses to edit tests to pass, and asks before changing anything.DeepSeek Harness「对抗式审查」Agent 预设:预设被审查对象是错的,每条结论必须给出权威依据与文件行号;严禁改测试让其通过,修改任何文件前先征求同意。