deepseek-harness/examples

Examples

Runnable demos (not workspaces) that showcase how the harness is wired. Each example is a thin leaf: a cordis.yml that picks the swappable backends (an LLM adapter, a bash executor), loads one app package, and may add optional product tools or demo-only mocks. The composition — the spine, the front-door cluster, and the boot glue — lives in the app packages (@deepseek-ai/dsh-stdio-demo, @deepseek-ai/dsh-cli-demo, @deepseek-ai/dsh-acp-demo) and the @deepseek-ai/dsh-agent-spine-demo bundle they share. There is no start.ts; the demo:* scripts invoke each app package's bin.

echo-agent

A mock model + echo tool on the stdio chat app — the all-mock skeleton. The leaf swaps dsh-stdio-demo's LLM backend to a local mock-echo adapter and adds a local echo tool. Demonstrates:

• A thin leaf cordis.yml loading the @deepseek-ai/dsh-stdio-demo app
• Registering a mock LlmAdapter (streaming scripted responses)
• Registering a tool via ctx.tools.register()
• "Swap the backend, keep the app" — the only difference from repl-agent is the adapter

Run with: pnpm run demo:echo. When prompted, type "echo " to trigger a tool call round-trip.

repl-agent

A coding agent with DeepSeek V4, the read/write/edit filesystem tools, the bash tool suite, subagent delegation, and the todo_write task tracker on the @deepseek-ai/dsh-stdio-demo app's readline front door.

Run with: pnpm run demo:repl (needs DEEPSEEK_API_KEY in the environment or a gitignored repo-root .env). See repl-agent/README.md for details.

Run the Code Mode overlay with pnpm run demo:code-mode, or pass acp for the ACP example. See the Code Mode example for its composition and a sample task.

headless-agent

A non-interactive agent demo that accepts one positional task, runs one complete model/tool turn on the @deepseek-ai/dsh-cli-demo app, persists a fresh session, prints text, json, or stream-json, and exits.

Run with: pnpm run demo:headless -- "task" (needs DEEPSEEK_API_KEY). See headless-agent/README.md for the output contract, safety boundaries, and snapshot suite.

tui-agent

The full-screen terminal sibling of repl-agent: it reuses the same coding backends and tools while forcing the shared terminal app to dsh-tui. It is the home of TUI PTY and snapshot scenarios.

Run with: pnpm run demo:tui (needs DEEPSEEK_API_KEY). See tui-agent/README.md for controls and composition.

jsonrpc-agent

An unattended coding agent driven through the Python SDK: JSON-RPC stdio, foreground-only bash, read / write / edit, one foreground subagent, todo_write, JSONL persistence, and compaction. It excludes terminal UI, stdout logging, approvals, skills, and background task controls. See jsonrpc-agent/README.md.

cordis-agent

The self-referential demo: the coding spine plus @deepseek-ai/dsh-tool-cordis, whose three tools (cordis_inspect / cordis_mount / cordis_unmount) let the agent inspect the live cordis runtime it runs inside, mount model-written plugins into it (an event listener, a brand-new tool for itself, or a service another mount injects), and dispose them again — all dynamic mounts grouped under one cordis-dynamic fiber subtree. The ctx.fs/ctx.web services ride along provider-only, as the capabilities those plugins build on.

Run with: pnpm run demo:cordis (needs DEEPSEEK_API_KEY). See cordis-agent/README.md for the staged demo script and the toolset Agent Note for the design and sandbox caveats.

acp-agent

An agent demo exposed as an Agent Client Protocol (ACP) server over JSON-RPC stdio, via the @deepseek-ai/dsh-acp-demo app — drive it from Zed or any other ACP client. It owns the ACP keyless snapshot suite.

Run with: pnpm run demo:acp (needs DEEPSEEK_API_KEY); pnpm run demo:code-mode acp boots the same server in Code Mode via the code-mode.cordis.yml overlay. See acp-agent/README.md for the Zed setup and the snapshot-test design.

The default cordis.yml composes @deepseek-ai/dsh-sandbox-local, @deepseek-ai/dsh-bash-sandbox, @deepseek-ai/dsh-user-approval, and @deepseek-ai/dsh-permission. A capable client gets one Permissions select: workspace-write confines bash to the configured workspace and asks before a wider retry, while danger-full-access removes file confinement and disables approval prompts. A denied command can therefore surface a one-shot session/request_permission prompt in the editor; "Allow once" runs exactly that retry under the requested wider mode.