블로그

A hands-on tutorial that builds a DeepSeek-Harness plugin: package.json, cordis.patch.yml, a registered tool, and installing it into a profile with dsh plugin add.

What Cordis is, where it came from, and how its Context, Service, and Fiber lifecycle primitives power DeepSeek-Harness's everything-is-a-plugin architecture.

Configure a DeepSeek API key in DeepSeek Harness via Settings → Models or settings.yaml, understand .credentials.yaml, and fix MISSING_CREDENTIAL errors.

How DeepSeek-Harness implements tools, commands, skills, MCP, hooks, and permissions as one mechanism, with the capability-seam pattern and packages panorama.

A reference for the dsh command line: the four entry modes, launcher flags, web and headless app parameters, plugin management, and every environment variable.

How DeepSeek Harness merges its four config layers, why a patch replaces a row instead of deep-merging it, and how to debug the result with --dump-config.

How a DeepSeek-Harness plugin author writes cordis.patch.yml: what insert, id, name, and config declare, and how bundle vs profile package.json fields differ.

Add OpenAI, Anthropic, Bedrock, or any OpenAI-compatible endpoint as a custom model provider in DeepSeek Harness, with the exact settings.yaml YAML fields.

A field-by-field guide to defineTool: parameters, output.schema, output.render, execute, the inject dependency it needs, and how Code Mode affects it.

30 short, fact-checked answers about DeepSeek Harness (dsh) — installation, cost, security, MCP, plugins, Claude Code comparisons, and the Python SDK.

A sourced roundup of what developers on X, Hacker News, Zhihu, V2EX, and LinuxDo said in DeepSeek Harness's first week — praise, complaints, and open disputes.

How to run DeepSeek Harness across a team: permission presets, sandbox backends, telemetry env vars, shared config layers, and a plugin approval policy.