deepseek-harness/packages/examples/stdio-demo

@deepseek-ai/dsh-stdio-demo

The terminal chat app: a Cordis app plugin that composes the default agent spine (@deepseek-ai/dsh-agent-spine-demo) with JSONL persistence, human interaction, a pre-created main agent, and a TTY-selected pi-tui/readline front door. Its bin boots a leaf cordis.yml.

It is the terminal counterpart to @deepseek-ai/dsh-acp-demo: both consume the same spine, while ACP reserves stdout for JSON-RPC and creates sessions from the client.

What it bakes in

A terminal chat always wants the same cluster, so the package owns it rather than trusting each leaf to re-wire it:

Plugin	Why it is here
@deepseek-ai/dsh-agent-spine-demo	the spine, pre-creating a main agent from this app's provider/model pair with process.cwd() as the fresh session cwd and carrying its persona
@deepseek-ai/dsh-session-persistence-jsonl	durable JSONL session log under persistenceRoot
@deepseek-ai/dsh-user-interaction	the human question/answer seam used by confirmation tools
@deepseek-ai/dsh-tool-ask-user	the model-facing ask_user_question tool
@cordisjs/plugin-logger-console	readline diagnostics for non-TTY operation; omitted from the fullscreen TUI path
@deepseek-ai/dsh-stdio	the line-oriented channel for pipes and automation, bound to the exact app-owned agent/session identity
@deepseek-ai/dsh-tui	the fullscreen interactive channel for TTY pairs, bound to the same exact identity

@cordisjs/plugin-hmr (the dev/demo edit-reload loop) is deliberately a leaf entry, NOT baked in here: it is a Loader-only, subprocess-only dev plugin — its constructor throws without node --expose-internals + a live loader, and the in-process test tier cannot even import it (so a package whose apply statically pulled it in could never carry the per-file coverage gate). Unlike the console logger, a stray hmr is not a stdout-purity footgun, so leaving it at the leaf costs no safety. The demo:echo / demo:repl leaves load it and pass --expose-internals.

The leaf cordis.yml supplies only the swappable backends — an LLM adapter (llm-deepseek for the real model, or the mock mock-llm for a demo) and a bash executor (bash-local) — hmr, plus this app's Config. The whole plugin tree a run loads is therefore: this app's cluster, the spine inside agent-spine-demo, hmr, and the two leaf backends.

Config

Key	Default	Routed to
provider	(required)	the pre-created main agent's registered provider route
model	(required)	the pre-created main agent's model
maxParallelToolCalls	agent-loop default	positive-integer concurrent tool-call cap shared by the bundled loop's agents; 1 is serial
persona	—	the deployment persona template (may reference {{provider}}/{{model}}/{{cwd}}), routed to dsh-system-prompt
toolOrder	—	explicit model-facing tool order (a name list with one '<unlisted-tools>' rest entry; absent — lexicographic; an unregistered name fails each turn at prompt assembly), routed to dsh-system-prompt
dshHome	$DSH_HOME or ~/.dsh	Harness home exposed to model bash and used by local skill discovery
tools	{ mode: 'native' }	tool-registry presentation config (native / code / both), routed through dsh-agent-spine-demo
skills	owner defaults	registry-cache, local-provider, and model-facing skill-tool config, routed through dsh-agent-spine-demo
toolBash	owner defaults	model-facing bash config routed through dsh-agent-spine-demo, including bash's producer-local enableRunInBackground
toolTasks	owner defaults	generic task_output wait bounds routed through dsh-agent-spine-demo
persistenceRoot	./.sessions	the JSONL backend's root directory
welcome	ready.	terminal banner / TUI subtitle
ui	{ mode: 'auto' }	terminal mode (auto / readline / tui) and nested TUI presentation config
resumeSessionId	—	resume a persisted session id instead of starting fresh (sourced from an env var in the leaf)

Fresh terminal sessions use the process launch directory as session.header.cwd and mint one combined main-session-<uuid> agent/session id, so durable restarts cannot collide. The app passes that exact opaque id to the config-created agent and selected UI before agent-core starts; this lets either front door observe agent-loop/config-start-failed, and an AgentLoop-only reload restores materialized history under the same id. Readline buffers startup input until agent/session-start; the TUI waits to enter fullscreen until the matching root appears. A resumed run binds both components to the exact resumeSessionId and keeps the persisted cwd.

The bin

dsh-stdio-demo [path-to-cordis.yml] (default ./cordis.yml) loads a gitignored .env from the cwd (DEEPSEEK_API_KEY / DEEPSEEK_BASE_URL), then drives the cordis Loader against the config and awaits the whole plugin tree before returning. Run it under node --expose-internals, or install the Loader's optional node-addon-require-builtin fallback, so the Loader can resolve the config's bare plugin specifiers (@deepseek-ai/dsh-*, npm packages). The demo:echo / demo:repl scripts use --expose-internals.

Example leaf cordis.yml

# A REPL agent demo: hmr + the DeepSeek adapter + local bash, then this app.
- id: hmr
  name: '@cordisjs/plugin-hmr'
  config:
    root: ['.']
- id: llm-deepseek
  name: '@deepseek-ai/dsh-llm-deepseek'
  config:
    apiKey: !!js process.env.DEEPSEEK_API_KEY
- id: bash
  name: '@deepseek-ai/dsh-bash-local'
  config:
    timeoutMs: 60000
- id: stdio-agent
  name: '@deepseek-ai/dsh-stdio-demo'
  config:
    provider: deepseek
    model: deepseek-v4-flash
    persona: 'You are a coding assistant powered by the {{model}} model.'
    ui:
      mode: auto

Swap llm-deepseek for a mock-llm leaf plugin and you have the echo demo — "swap the backend, keep the app".

Model Experience

Composed terminal agent request

What the model sees

Through dsh-agent-spine-demo, the main agent receives the harness identity, configured persona, skill catalog, and visible tools; this app also composes the generated ask_user_question schema. Each terminal submission becomes a user message; submissions made while the agent runs steer the active turn.

Token effect

Child prompt and schema costs repeat per request; user input and tool history grow until compaction. Terminal banners, logger output, cards, and rendered transcripts add zero model tokens.

KV Cache effect

User and tool history is append-only while the composed prompt, schemas, child model route, and session prefix remain fixed. A composition change or compaction may invalidate reuse from its first changed token; terminal rendering has no cache effect.

Human-answer result

What the model sees

Through dsh-tool-ask-user, successful terminal answers use that package's exact compact JSON shape. Interruption becomes exactly Error: ask_user_question was interrupted before the user answered; a closed stdin becomes Error: ask_user_question cannot be answered because stdin is closed.

Token effect

Only a completed or failed tool call adds retained result tokens; prompts printed while waiting are terminal-only.

KV Cache effect

Append-only; newly visible content follows the reusable request prefix and does not invalidate existing KV-cache entries.

Known Limitations and Deferred Work

• One pre-created main agent drives the selected terminal UI — there is no multi-session or concurrent-agent surface in this app; a run is one conversation.
• The front-door cluster is fixed in code — the JSONL persistence backend and the ask-user tooling are baked; a different composition is a leaf-level sibling entry or another app package.
• The question tool is not an approval answerer — this app mounts user-interaction and ask_user_question, but not ctx.approval; a tools/pre-execute ask therefore fails closed unless the leaf composes an approval service and terminal answerer.