Replace the Task-backed continuation manager with one durable Session plus at
most one process-local Activation — a residency epoch for a reconstructed child
Agent, not a request, result, cancellation, or Task boundary. The manager owns
activation admission, authority, the live ownership graph, cold resume, and
child-first disposal; the Agent inbox is the only turn FIFO.
- startContinuable() is async and returns { childId, messageId } at inbox
acceptance; followup() takes a SubagentAuthority and returns AgentMessageId.
- SubagentProvider.resume?(), SubagentProviderResumeRequest, SubagentRun.steer?(),
SubagentProviderStartRequest and SubagentContinuation are deleted;
prepareContinuable?() is the continuable-creation capability.
- Cold resume calls ctx.agents.resume() from the manager through a private
activation-owner scope, never dispatching through a provider.
- Extract shared child composition, descriptor seeding, depth accounting, and
one-shot run settlement so the manager and one-shot driver keep one home
per fact.
Tests and docs follow in subsequent commits.
Implement the continuable background subagents RFC: a durable child
session with a series of Task-backed activations, each disposing its
run before the Task settles.
- dsh-subagent: rename SubagentRun.sendMessage to strict steer, drop
run-level resume, add SubagentProvider.resume dispatch via
SubagentService.resume, the continuation start field, and the
versioned model-hidden subagent/descriptor session event.
- dsh-subagent-inprocess/-spawn/-fork: publish the control-allocated
child id, append the descriptor inside the initial turn, implement
cold resume from the child's own transcript under the live parent
scope, and strict running-only steer.
- dsh-subagent-control (new): SubagentControlService owning stable
child ids, descriptor snapshot/fold/authorization, Task-backed
activation with settle-then-dispose ordering, the process-local
active-run association, and steer-or-resume sendMessage routing.
- dsh-tool-subagent: background route branches on the provider's
resume capability (continuable via the control service; one-shot
task for ACP), returning both child and task ids.
- dsh-tool-subagent-control (new): the globally named send_message
tool rendering steered/started routes.
Keyless coverage spans Task ownership and disposal ordering, running
delivery, cold follow-up, descriptor rejection and rollback, known-id
reconstruction, kill during lookup, admission races, and a new
subagent-continuable ACP snapshot scenario.
SubagentStartRequest gains persona (capability-gated like toolFilter); the
in-process driver composes the child's scoped world in the factory's setup
window — persona as a scoped shadowing deployment:persona section,
toolFilter as a scoped tools.restrict() (loud unknown-name validation),
outputSchema as the scoped structured runtime. spawn/fork now advertise
every start-time capability; ACP stays all-false. A parent-scope teardown
effect links each child to its parent through the memoized handle, so a
disposed parent reaches its whole subtree even if the delegating tool's
finally never runs; subagent/start|end dispatch in the delegating parent's
scope.
structured.ts loses the placeholder schema, the final-assembly swap/strip,
the refcounted root runtime, and the WeakMap state: each child registers
its OWN capture tool (real schema), instruction section, and enforcement
listeners on child.ctx, riding the child's fiber. The commit listener is
call-keyed (a stale stage from a short-circuited post-execute chain is
dropped, never promoted on a later call), and one scoped prepend re-assert
listener preserves the final-assembly guarantee against a stripping global
listener.
tool-subagent gains persona/toolFilter/maxDepth passthrough config —
deny-listing the delegation tool (or maxDepth) is how a deployment bounds
recursion; the omitted-toolFilter schema key is forced absent (a
materialized {} would mean an empty allow-list, i.e. deny-everything).
The Codex simplification concern plus the duplication comment on the spawn
apply, resolved by deletion: the backend-lifetime holds are gone, so the
runtime registers at the first structured run and disposes when the last
settles — a deployment that never passes outputSchema carries no always-on
global state, and there is no per-backend acquisition block left to extract.
The driver spec now drives an INLINE spawn-shaped provider over
startInProcessRun, which removes the spawn/fork devDependencies (the
test-only workspace cycle); plugin-level structured coverage moves to the
backends' own specs (capture through the shipped plugin, mid-run backend
unload, seeded fork capture). tools.md, the driver README, and both backend
READMEs describe the run-scoped lifetime; the module-graph regenerates
without the cycle edges.
Two human review directives:
- No re-prompt. A structured child that finishes a turn cleanly without
calling structured_output settles error to the parent immediately —
readResult already carried that mapping; the nudge loop only delayed it.
Deletes the loop, its cancellation-window guard, STRUCTURED_OUTPUT_NUDGE,
and the structuredNudgeRetries Config on both backends.
- FIXME in the structured module doc: per-agent/per-session tool registry and
prompt assembly would dissolve the final-assembly enforcement dance (the
placeholder tool, the swap, the strip, the global-registration lifetime).
Carved out of #170 per review feedback — the foundation the workflow tool
builds on, now standing alone on master:
- dsh-tools: the structured-output JSON Schema subset (StructuredOutputSchema,
assertSupportedOutputSchema, validateStructuredValue) — rejects loud outside
the enforced subset, listing every violation
- dsh-subagent: SubagentStartRequest.outputSchema / SubagentResult.structured
become a real capability; the service rejects a schema'd request whose
provider lacks it
- dsh-subagent-inprocess: the shared structured runtime — one global
structured_output capture tool, a prepend final-assembly listener that
strips the placeholder for plain agents and swaps in the run's own schema
(plus the calling instruction as a trailing section) for structured
children, an agent/turn-continuation veto once captured, and the
capture/nudge loop in the run driver (structuredNudgeRetries, cancellation
honored mid-nudge); lifetime refcounted by backends and live runs
- subagent-spawn / subagent-fork flip outputSchema: true
One deliberate divergence from the #170 revision: the backends do NOT add
'tools' to their plugin inject. Doing so deferred their apply past the todo
plugin, and the delegation tool mirrors provider lifecycle — so the
model-visible tool order of every existing prompt changed, invalidating every
recorded snapshot fixture. The runtime now gates its capture-tool registration
on tools availability itself (sync when live, a scoped inject fiber when the
Loader starts the backend first), keeping this PR byte-invisible to existing
transcripts: all 35 snapshot scenarios pass against master's fixtures
unchanged.
One principle: every fact in the assembled prompt has exactly one owner.
- dsh-system-prompt: merge-extensible AssembleContext on assemble();
a variable(name, provider) registry; {{name}} interpolation in
renderPrompt, strict (unknown/valueless/malformed references throw);
duplicate section and variable names rejected; assembly carries
resolved section text + variables through the assemble waterfall.
- dsh-agent declares AssembleContext.agent; dsh-agent-loop registers
the agent:persona section (order 0 - identity renders before tool
guidance) and the model/cwd variables, and drops its string join:
renderPrompt(assembly) IS the full prompt.
- Tool guidance moves to its owners: descriptions carry per-tool
semantics; sections only cross-call habits (tool:bash exit-code
habit at order 105; read's not-shell nudge). todo/subagent need no
section - their descriptions already carry the contract.
- SubagentProvider.inheritsParentContext (spawn/acp false, fork true);
dsh-tool-subagent derives truthful per-provider wording and resolves
the provider at load (backend must be listed first).
- Example personas shrink to identity + behavior with {{model}} (and
{{cwd}} in the ACP tree); the welcome banner stops enumerating tools.
RFC: docs/rfc/implemented/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.md
The shared run driver lived inside dsh-subagent-spawn, so the spawn package
carried fork-aware seeding logic and dsh-subagent-fork depended backward on
dsh-subagent-spawn — the two in-process backends were not independent.
Move the driver (startInProcessRun, depthOf, SubagentDepthError,
InProcessRunOptions) into a new pure-library package
@deepseek-ai/dsh-subagent-inprocess that registers nothing. spawn and fork now
both depend only on that driver and neither knows about the other; spawn no
longer re-exports it and fork no longer imports from spawn.
Also wire BOTH backends in examples/coding-agent/cordis.yml (config-only): load
dsh-subagent-spawn + dsh-subagent-fork + two dsh-tool-subagent instances with
distinct toolNames (subagent → spawn, subagent_fork → fork), demonstrating that
exposing multiple transports needs no code change.
The second PR of the subagent seam: the two in-process backends that run a
child agent on the same cordis context, reusing the agent factory's quiescent
AgentHandle teardown. Both register on ctx.subagents (PR1's named-provider
registry) and share one run driver.
- dsh-subagent-spawn: a FRESH child via ctx.agents.create — own session, the
parent's model by default (overridable), zero inherited conversation. Also
exports the shared in-process run driver (startInProcessRun): mint ids, stamp
cwd/parentSession-lineage/depth, drive the one-shot (send → whenIdle), read
the last assistant/message + turn/end reason, dispose to quiescence.
- dsh-subagent-fork: a child SEEDED with the parent's balanced completed-turn
prefix (the log up to and including its last turn/end), so the child inherits
context. The in-flight unbalanced turn is excluded — a raw seed would fail the
invariants replay. Proven: a regression test goes red if the boundary seeds
the open turn.
- Seam extension: CreateAgentOptions.seed, threaded through AgentLoop.createAgent
→ ctx.sessions.prepare({ seed }) (the primitive resume already used). This is
the fork-lineage path the TODO(sub-agents) markers anticipated.
- Depth: a merge-extensible AgentOptions.subagentDepth (0 top-level, parent+1 for
a child); the depthLimit capability refuses a spawn past request.maxDepth.
Tests: real-loop unit tests for both backends (mock MODEL only, real loop +
invariants), a multi-subagent test (one parent drives a fork AND a spawn child
then keeps working), and a with-key e2e (a real parent delegates via the
`subagent` tool to a real child that writes a file on disk — world-verified).
100% per-file coverage. The coding-agent demo wires the spawn backend + tool.
Snapshot coverage of nested agents is deferred to a stacked follow-up
(TODO(subagent-snapshots)): dsh-llm-replay is a single global positional cursor
that cannot route calls to a parent vs. a child on one context. Recorded in the
RFC's deferrals and a new AGENTS.md rule: designing a subsystem must design its
test infrastructure END TO END up front, verifying the snapshot/e2e harness can
express the new shape — a gap this plan hit.
