Tianyi Cui
3f83a4ee96
Review round 2 (tianyicui inline comments): - dsh-system-prompt itself registers the harness:identity (-100) and deployment:persona (0) sections — they must survive a swapped loop plugin, so they leave dsh-agent-loop; the persona text is the plugin's own validated 'persona' config. The model/cwd variables STAY on the loop: runtime facts of the agents it drives. - AgentOptions.systemPrompt is deleted with all its forwarding plumbing: the app configs' systemPrompt keys become 'persona' routed through dsh-agent-core (schema = z.intersect of the owners'), the ACP bridge and tool-subagent stop carrying persona configuration, and subagent children now render the deployment persona like every other agent. - Example personas drop transport/interface trivia (ACP, CLI) — facts irrelevant to the model. - Root CONTEXT.md removed (not idiomatic); its persona definition was wrong under the new ownership anyway. - Docs, READMEs, the prompt-variables RFC, and generated catalogs updated; new loop test pins the assemble-waterfall escape valve (an emptied assembly sends NO system field).
echo-agent
Runnable demo: stdin chat with a scripted mock model and an echo tool. The all-mock skeleton — "swap the backend, keep the app".
What it shows
This example is just a leaf cordis.yml: it loads the @deepseek-ai/dsh-stdio-agent app (which bundles the whole @deepseek-ai/dsh-agent-core spine, the console logger, JSONL persistence, the readline UI, and a pre-created main agent), and swaps in two example-local backends plus hmr:
mock-llm.ts— a mockLlmAdapterthat streams scripted responses and calls theechotool when the user types "echo ". Registered withctx.llm.registerAdapter(['mock-echo'], …).echo-tool.ts— a tool registered viactx.tools.register(defineTool(…))with typedexecuteargs; echoes text back uppercased.
Swapping mock-llm for the real llm-deepseek adapter is all that separates this from coding-agent — the same app, a different backend.
Plugin files
| File | Role | Key patterns demonstrated |
|---|---|---|
src/mock-llm.ts |
LlmAdapter registration |
ctx.llm.registerAdapter(['mock-echo'], …), streaming chunks with the proper block-start/block-end protocol |
src/echo-tool.ts |
Tool registration | ctx.tools.register(defineTool(…)) with typed execute args, returning ContentBlock[] |
cordis.yml |
Leaf wiring | the two backends + hmr + one @deepseek-ai/dsh-stdio-agent entry carrying the app config |
The spine, UI, persistence, and boot glue all live in @deepseek-ai/dsh-stdio-agent and the bundle it loads — this folder holds only the demo-specific mocks and the leaf wiring.
Run
pnpm run demo:echo
# or:
node --expose-internals --import tsx packages/ui/stdio-agent/src/bin.ts examples/echo-agent/cordis.yml
Type a message and press Enter. "echo " triggers a tool call round-trip (the mock model requests the echo tool, which echoes the text uppercased, and the next model step acknowledges it).
The session is persisted under .sessions/ relative to the directory you launch the demo from. pnpm run demo:echo runs from the repo root, so the logs land in <repo-root>/.sessions/ (a session with no cwd goes in the _no-cwd/ bucket, one .jsonl log per session). Clean up with: rm -rf .sessions