Files
deepseek-harness/packages
Tianyi Cui 1a81f2cccd Add subagent capability seam: interface, mock backend, model-facing tool
Introduce the `packages/subagent/` group and the abstract subagent seam — an
agent delegating to a child agent — as a named-provider registry (`ctx.subagents`),
unlike the single-implementation bash seam, so multiple transports (in-process,
ACP, future A2A) coexist. This first PR lands the interface, a scripted test
backend, and the model-facing tool, validated through the real cordis load path.

- dsh-subagent: SubagentService registry + SubagentProvider/SubagentRun
  vocabulary + subagent/start|end events. Start-time capabilities (outputSchema,
  depthLimit, toolFilter) are checked pre-start and rejected loud; runtime
  capabilities (sendMessage, resume) are optional methods on SubagentRun.
- dsh-subagent-mock (support): scripted provider for keyless, deterministic
  tests through the real Loader/export path.
- dsh-tool-subagent: the model-facing `subagent` tool, config-bound to one
  provider; synchronous collect with try/finally dispose, signal->cancel
  bridging, and non-completed-stop-reason -> isError mapping.
- Proposed RFC documenting the seam, the fork-vs-spawn-as-separate-backends
  decision, own-session isolation, synchronous-collect scope, and the deferral
  of background/poll/spill to a future unification with bash.
- Wire the new group into tsconfigs, build refs, package hierarchy docs, the
  module graph, and the cordis catalog.

RFC: docs/rfc/proposed/feature/2026-06-21-subagent-capability-seam.md
2026-06-21 22:31:56 +08:00
..

Packages

Harness packages, all under the @deepseek-ai/dsh-* scope. Each package is a Cordis plugin (microkernel-style): it exports either a default Service subclass or a functional plugin that gets registered via ctx.plugin(), declares its ctx key/events where applicable through declaration merging, and exposes extension points through ctx.effect(), ctx.on(), and ctx.waterfall().

Hierarchy

Packages are grouped by modular role at packages/<group>/<pkg>/. The group directory is a pure container (no package.json of its own); the package name stays @deepseek-ai/dsh-<pkg> regardless of group. Each group has a README.md describing its role and whether it is product or support infrastructure.

Group Role Release expectation
core/ Product API spine: session, system-prompt, tools, agent, and the concrete loop Product — stable surface
llm/ LLM capability family: the abstract service + provider adapters Product — stable surface
bash/ Bash capability family: the executor seam, a local impl, and the model-facing tool Product — stable surface
subagent/ Subagent capability family: the provider-registry seam and the model-facing delegation tool Product — stable surface
session-persistence/ Persistence capability family: the seam + JSONL/SQLite backends Product — stable surface
ui/ Editor/client integration surfaces (the ACP bridge) Product — stable surface
support/ Dev/test/example infrastructure (invariants, stdio UI, replay adapter) Support — lower compatibility expectations
util/ Low-level zero-dependency utilities shared across groups (the Branded<B> primitive) Support — small, stable, harness-dep-free

The split is the point: a package's group says whether it is part of the product API or support/test/example infrastructure, so release and removal decisions do not have to treat every package as an equal public contract. New packages join an existing group; adding a new top-level group is a deliberate act (extend the group READMEs and the hierarchy docs).

Dependency graph

dsh-brand         (no harness deps — type-only Branded<B> primitive)
dsh-llm          ← dsh-brand                       (vocabulary; brands CallId)
dsh-bash          ← dsh-brand                       (abstract executor seam; brands BashTaskId/OwnerToken)
dsh-session       ← dsh-llm, dsh-brand
dsh-system-prompt ← dsh-llm
dsh-agent         ← dsh-llm, dsh-session, dsh-brand
dsh-tools         ← dsh-llm, dsh-system-prompt, dsh-agent
dsh-bash-local    ← dsh-bash                       (BashExecutor impl)
dsh-tool-bash     ← dsh-bash, dsh-tools            (bash tool schemas)
dsh-llm-deepseek  ← dsh-llm                        (DeepSeek adapter)
dsh-llm-pi-ai     ← dsh-llm                        (pi-ai-backed adapter)
dsh-agent-loop    ← dsh-llm, dsh-session, dsh-system-prompt, dsh-tools, dsh-agent
dsh-invariants    ← dsh-llm, dsh-session, dsh-agent (dev-mode contract checks)
dsh-acp           ← dsh-agent, dsh-llm, dsh-session, dsh-session-persistence  (ACP JSON-RPC bridge)
dsh-ui-stdio      ← dsh-agent, dsh-llm, dsh-session (stdio readline UI plugin)
dsh-llm-replay    ← dsh-llm, dsh-session            (record/replay adapter for keyless snapshot tests)
dsh-subagent      ← dsh-agent, dsh-llm, dsh-tools    (abstract subagent provider-registry seam)
dsh-subagent-mock ← dsh-subagent                     (scripted provider for tests)
dsh-tool-subagent ← dsh-subagent, dsh-tools, dsh-agent (model-facing delegation tool)
dsh-agent-core    ← timer, dsh-llm, dsh-session, dsh-system-prompt, dsh-tools, dsh-agent, dsh-invariants, dsh-tool-bash, dsh-agent-loop  (the providerless spine, as one bundle plugin)
dsh-stdio-agent   ← dsh-agent-core, dsh-ui-stdio, dsh-session-persistence-jsonl, dsh-agent, dsh-session  (stdio chat APP + bin)
dsh-acp-agent     ← dsh-agent-core, dsh-acp, dsh-session-persistence-jsonl     (ACP server APP + bin)

The rule: extension plugins depend on interfaces, never on the concrete loop. dsh-agent-loop is swappable — UI/hook/tool plugins keep working against the dsh-agent vocabulary if the loop is replaced. The sanctioned exception is a composition/bundle package like dsh-agent-core, whose whole job is to assemble the concrete spine: it depends on dsh-agent-loop (and the other concrete spine plugins) on purpose. The rule constrains plugins that EXTEND the system, not the bundle that COMPOSES it — swapping the loop means shipping a different bundle, not rewiring every extension. A swappable capability splits into interface / implementation / consumer packages (the bash trio is the template — see capability seams).

What goes where

Package Group Role ctx key
llm/ llm Abstract LLM service + content-block vocabulary + chunk assembler ctx.llm
session/ core Event-sourced session log + in-memory store ctx.sessions
system-prompt/ core Prompt-section + tool-schema assembly registry ctx.systemPrompt
tools/ core Tool registry + tools/execute waterfall ctx.tools
agent/ core Agent interface, registry, agent/* event vocabulary ctx.agents
agent-loop/ core THE concrete loop plugin: ReactLoopAgent + the loop driver ctx.agentLoop
agent-core/ core Bundle plugin: the providerless/executor-less/UI-less spine as code (forwards agent-loop's agents) (loads the spine)
bash/ bash Abstract bash executor seam (interface + vocabulary) ctx.bash
bash-local/ bash Local-subprocess BashExecutor implementation (registers ctx.bash)
tool-bash/ bash Model-facing bash/bash_output/bash_kill tool schemas (registers on ctx.tools)
llm-deepseek/ llm DeepSeek API adapter (hand-rolled fetch/SSE) (registers on ctx.llm)
llm-pi-ai/ llm DeepSeek adapter via @earendil-works/pi-ai (design twin) (registers on ctx.llm)
session-persistence/ session-persistence Persistence seam + write coordinator ctx.sessionPersistence
session-persistence-jsonl/ session-persistence JSONL-sidecar persistence backend (registers ctx.sessionPersistence)
session-persistence-sqlite/ session-persistence SQLite persistence backend (registers ctx.sessionPersistence)
invariants/ support Dev-mode event-contract invariants + session-log freeze (listens on session/*, agent/*)
acp/ ui Agent Client Protocol bridge: serves the agent to an ACP editor over JSON-RPC stdio (drives ctx.agents/ctx.sessions)
stdio-agent/ ui Terminal stdio chat APP: agent-core spine + console logger + readline UI + a pre-created main agent, with a bin (composition + bin)
acp-agent/ ui ACP server APP: agent-core spine + JSONL persistence + the acp bridge (no stdout logger), with a bin (composition + bin)
ui-stdio/ support Minimal stdio (readline) UI plugin: renders agent/* events, feeds stdin lines to the agent (drives ctx.agents)
llm-replay/ support Record/replay adapter: short-circuits llm/stream with chunks from a recorded session JSONL (keyless snapshot tests) (listens on llm/stream)
subagent/ subagent Abstract subagent seam: named-provider registry for delegating to child agents ctx.subagents
subagent-mock/ support Scripted SubagentProvider for testing the seam through the real load path (registers on ctx.subagents)
tool-subagent/ subagent Model-facing subagent delegation tool over ctx.subagents (registers on ctx.tools)
brand/ util Type-only Branded<B> nominal-typing primitive (no runtime code, no harness deps) (none — type-only)

Each package has its own README.md with purpose, service API, events, extension points, and deliberate non-goals (TODOs).

Conventions (applied across all harness packages)

  • Registrations are effects: every contribution (adapter, tool, section, agent, event listener) goes through ctx.effect() / ctx.on(), so disposal and HMR clean up automatically. Every register() returns the disposer.
  • Declaration merging for events and ctx: services declare their events in declare module 'cordis' { interface Events { ... } } and their ctx key in interface Context.
  • Waterfall semantics: ctx.waterfall listeners receive (...args, next) and MUST call next() to delegate; returning without it short-circuits (the veto mechanism).
  • Extensible unions: ContentBlockMap, MessageSourceMap, FinishReasonMap, TurnTriggerMap, TurnEndReasonMap, and SessionEventMap use the merge-extensible-map pattern so plugins can add variants via declaration merging.
  • ESM everywhere; imports use package names across package boundaries, .ts extensions within a package.
  • Tests: vitest, colocated under packages/<group>/<pkg>/tests/*.spec.ts. Every registry needs an HMR-safety test. Err on the side of more tests.