The group's convention is package suffix == provider default
(subagent-acp/'acp', subagent-spawn/'spawn', subagent-fork/'fork'), and the
provider default became dsh-sdk in the last review round — so the package
follows: @deepseek-ai/dsh-subagent-dsh-sdk at
packages/subagent/subagent-dsh-sdk, plugin name subagent-dsh-sdk,
diagnostics prefixed subagent-dsh-sdk:. The dsh echo has precedent
(dsh-llm-deepseek). Directory, fixture path, knip/tsconfig/examples
registrations, catalogs, READMEs (en+zh), and the Agent Note follow; the
sdk-client dispose ladder moves to its own module (src/dispose.ts) with the
deterministic FakeChild tier tests restored alongside it.
Master's #660 replaced dsh-subagent-subprocess with the dsh-subprocess
capability seam (ctx.subprocess + scrubbedParentEnv, tree-scoped teardown)
and moved subagent-acp onto it. Convergence for this branch's packages:
- The shared out-of-process provider vocabulary this branch had grown in
the deleted library (NO_START_CAPABILITIES, assertPositiveFinite, cwd
resolution, settleRunResult, subprocessRunHandle) moves into the subagent
seam package as out-of-process.ts — it enforces subagent-seam contracts,
not process mechanics, and both out-of-process backends now import it
from there (subagent-acp keeps master's shape otherwise).
- subagent-sdk spawns THROUGH the SDK client (the subprocess README's
documented exception for SDK-managed transports) and now applies the
seam's scrubbedParentEnv() + explicit-env merge in place of the deleted
buildChildEnv.
- sdk-client inlines the EOF→SIGTERM→SIGKILL ladder as private helpers (it
runs outside any harness context, so it cannot ride ctx.subprocess).
- The child harness fixture gains the now-required dsh-subprocess-local
entry for bash-local; the fixture cordis.yml keeps exercising the
shipped provider default.
Resolutions: regenerate the conflicted generated docs (cordis services
catalog, event-producer-consumer, module-graph); take master's
packages/README pair and re-insert the telemetry row on both sides;
re-record the README and session-doc translation pairs.
Keep restore-only and self-hosted stores outside pnpm/action-setup's replaceable install directory, resolving the configured path for every paired cache. Pin Yarn 4 for the generated-project e2e so runner images without Corepack do not fall back to Yarn Classic, and update the bilingual decision record.
The duplication gate flagged three ACP/SDK clones; the shared halves move
into dsh-subagent-subprocess as provider.ts: NO_START_CAPABILITIES (frozen
all-false advertisement), assertPositiveFinite (prefix-parameterized timing
validation), settleRunResult (never-reject result settlement with contained
onError sink and listener hygiene), and subprocessRunHandle (idempotent
dispose publication). Both backends now compose these; the previously
unreachable cancelled-rejection branch is directly unit-tested at the
library level instead of v8-ignored in each backend. knip learns the
subagent-sdk workspace (e2e entry outside vitest unit includes).
- examples/jsonrpc-agent gains its first snapshot suite (sdk.snapshot.ts):
the real dsh-jsonrpc-agent runtime driven through the real dsh-sdk-client,
keyless llm-replay behind a new cordis.snapshot.yml overlay; three recorded
scenarios (text turn, bash tool, spawn subagent) pin the notification
stream, the SDK turn result, and the persisted parent+child session logs.
- Bilingual READMEs for dsh-sdk-protocol / dsh-sdk-client / dsh-subagent-sdk;
sdk/ and subagent/ group tables extended; dsh-jsonrpc README points at the
extracted protocol package; Agent Note (en+zh) owns the decision.
- The proposed make-jsonrpc-directional note is updated for the transport's
new home and its second (client) consumer.
- Model Experience sentence allowlist entries for the two client-side
packages; module graph + config catalog regenerated; i18n pairings
recorded. doc-sync passes 24/24.
SubprocessHandle loses dispose(graces) and SubprocessDisposeGraces: the
stdin-EOF→SIGTERM→SIGKILL sequence is teardown POLICY encoding one
consumer's cooperation shape, not process vocabulary — the seam keeps
kill/terminate/waitForExit, and waitForExit(signal?) is the quiescence
probe a consumer ladder needs. dsh-subagent-acp owns disposeAcpChild()
over those public verbs (tier tests move into its suite; a never-exits
stub pins the fail-loud path); dsh-subprocess-local sheds the ladder,
its deadline import, and the dsh-timeout dependency. Every future
backend now owes four verbs and no teardown policy. New bilingual
ladder-ownership Agent Note records the decision; catalogs regenerated.
- @deepseek-ai/dsh-sdk-protocol: extract the line transport from dsh-jsonrpc
and name the request/result/notification wire types both ends share; error
responses preserve wire code/data via JsonRpcResponseError.
- @deepseek-ai/dsh-sdk-client: TypeScript twin of the Python SDK — spawns the
dsh-jsonrpc-agent runtime as a subprocess, drives stdio JSON-RPC turns
(DeepSeekHarness high-level API + HarnessClient protocol client), scopes
notifications to session trees client-side, and reaps the child through the
shared subprocess dispose ladder.
- @deepseek-ai/dsh-subagent-sdk: out-of-process subagent backend driving a
child harness runtime through the TS SDK; shares cwd resolution with
subagent-acp via new dsh-subagent-subprocess cwd helpers.
- Keyless unit suites drive real subprocesses (scripted fake runtime peer);
100% per-file coverage on all touched packages.
Implements the turndown Agent Note from the NIH dependency audit (full
variant, not the minimal entities-only fallback): dsh-tool-web's fetch
rendering now converts HTML through turndown + @joplin/turndown-plugin-gfm
(atx headings, fenced code, dash bullets, GFM tables/strikethrough) over
the real domino DOM, with script/style/noscript removed wholesale. The
hand-rolled ~86-line regex converter html.ts and its entity tables are
deleted; renderBody wraps the conversion in try/catch falling back to
the raw HTML body, because turndown's recursive DOM walk overflows with
a RangeError on pathological nesting (measured: 4k levels on the main
thread, 8k in a worker) where the regex version could never throw.
Closure weight, measured: tool-web IS in the single-exe runtime closure,
and the exe asset globs would pack ~7.9 MB of the three new packages —
but ~6 MB of that is domino's test corpus, with runtime lib/ at ~550 KB
against a ~174 MB artifact (<0.5% either way), so the swap wins.
Per testing policy the previously-missing keyless web_fetch snapshot
ships in the same change: the acp-agent `web-fetch` scenario boots a new
web.cordis.yml overlay (web seam + real dsh-web-fetch-local provider +
tool-web fetch-only + a loopback HTTP fixture server on a fixed port
serving deterministic HTML with entities, a GFM table, and nesting), so
recording and keyless replay both drive the real HTTP fetch and real
conversion end to end; the scenario pins the new `web` header class.
The Agent Note moves proposed -> implemented and is rewritten per the
lifecycle contract (Decision/Consequences/Testing, closure verdict and
alternatives recorded); tool-web and acp-agent READMEs updated in both
languages and pairs re-recorded.
Implements the approved simplification Agent Note: sse.ts now pipes the
response body through TextDecoderStream and EventSourceParserStream
(eventsource-parser/stream) and keeps only the DeepSeek protocol shim —
yield each event's data, terminate on [DONE], throw
LlmError('STREAM_CLOSED') on EOF without the sentinel. The SSE
spec-conformance tests are deleted; sse.spec.ts pins only the
[DONE]/STREAM_CLOSED/EOF contract, including the new spec-strict verdict
that an unterminated trailing event is truncation (the old parser
flushed it — a robustness nicety no real provider shape needs).
eventsource-parser@^3.1.0 becomes llm-deepseek's second runtime
dependency (already in the lockfile transitively via the MCP SDK).
Docs: the Agent Note moves proposed/ → implemented/ and is rewritten per
the lifecycle contract; the rejected NIH roll-up note's inbound links
follow. The twin-adapters note, dsh-llm LlmAdapter JSDoc (and its
type-equiv fences), cookbook, group/package READMEs, root AGENTS.md
layout line, sdk-helper comments, and the regenerated config catalog
drop the "hand-rolled fetch + SSE" claim in both languages; all eight
touched pairs re-recorded.
Implements the execa Agent Note's four sub-changes:
- execa (root devDep + loader-smoke dep) replaces the hand-rolled
spawn-collect-timeout choreography in loader-smoke, apps/cli and
cli-demo/acp-demo built-bin e2e, lsp-local and code-runtime-worker
built-lib e2e, the tui pty-harness outer collector, the jsonrpc
keyless smoke, and crash-recovery's child spawn. Genuinely custom
parts stay custom: cli-demo's interrupt-on-marker, jsonrpc's
line-predicate protocol driving, crash-recovery's SIGKILL-at-failpoint.
The two loader-smoke /* v8 ignore */ OS-error branches are gone.
- llm-mock-server CLI tokenizes via node:util parseArgs; numeric
coercion/bounds/cross-option constraints stay manual; pinned
error-message tests updated to the parseArgs texts.
- both loadRootEnv copies in apps/web/tests are deleted: the owning
vitest configs (web unconditionally, snapshot in record mode)
already load the repo-root .env before these files run.
- the four poll loops (acp-snapshot harness waits + crash-recovery
waitForFile) ride vi.waitFor with explicit {interval, timeout}.
Resolutions: regenerate docs/capability-seams.md (the only conflict);
master widened the bilingual pairing scope to package READMEs, so this
merge adds the Chinese counterparts and i18n records for the three
telemetry READMEs and patches packages/README.zh.md with the telemetry
group row.
(A1) terminate()/dispose()/service teardown keyed on direct-child settlement
could leak a TERM-trapping descendant that outlived the leader (Codex
reproduced it with a disowned trap-SIGTERM helper). kill()/terminate() now
gate on tree liveness instead of outcome settlement; the SIGKILL escalation
timer survives settle (unref'd, re-probing the tree); dispose's tier
quiescence is whole-tree exit via a bounded waitForExit; the service's live
set releases handles only when their tree is gone, and its teardown awaits
tree exit. Three new suites pin the survivor scenarios end to end
(terminate, dispose, service teardown).
(A2) the escalation branch is now real tested behavior — its ignore is gone;
the one remaining signalTree guard ignore states why it is unreachable
through the handle verbs.
(A3) docs contradictions fixed: the impl README's stale POSIX-only bullet
now states the contained best-effort Windows tree story; the lsp-local
README no longer claims taskkill failures stay visible (containment + the
tree-liveness wait is the actual contract); the architecture tables (en+zh)
list all three consumer families.
(B1) OutputCollector keeps a byte-exact tail across uneven chunk boundaries
(trim the head chunk instead of dropping it whole) — the LSP diagnostic-tail
contract; pinned by a cross-chunk test.
(B2) the subagent-acp coverage ignore is narrowed to exactly the
never-settling success arm.
Review direction (tianyicui, PR #660): in a stacked PR, change all other
process-running places to use the new service.
- lsp-local: LspConnection spawns through ctx.subprocess (piped protocol
streams + a no-spill collected stderr tail); its private process-tree
helpers (POSIX group signalling, Windows taskkill, liveness polling) are
deleted in favor of the seam's handle verbs, and its buildChildEnv now
rides scrubbedParentEnv (LSP children also stop inheriting stale DSH_*).
The plugin injects 'subprocess'; compositions/tests mount
dsh-subprocess-local.
- subagent-acp: the ACP child spawns through the seam (piped ndjson streams,
inherited stderr); spawn failure surfaces through done-rejection into the
same startup race; disposal is handle.dispose with the plugin's configured
graces. dsh-subagent-subprocess is DELETED — its dispose ladder and scrub
are the seam's, and the isolated-config-dir helper had no consumer.
- mcp-client, pty-local, sdk-helper: adopt scrubbedParentEnv as the one
scrub definition (their spawns stay put by ownership: the MCP SDK and
node-pty own those calls; the SDK wizard runs outside any composition).
- Coverage: per-file 100% over every touched src file, with each v8 ignore
carrying a platform or contract reason; new suites cover stdio
dispositions, the dispose ladder tiers, injected-win32 tree semantics,
waitForExit, settled-kill/terminate no-ops, and spawn-failure disposal.
- Docs: consumer-migration Agent Note (en; zh follows in this PR), seam note
updated in place, subprocess.md rewritten for the reshaped vocabulary
(type-equiv re-registered), READMEs and SERVICE_ROLES updated, taskkill
added to knip ignoreBinaries.
ui-sidebar shell landed on master (#643: geometry-only shell, the
browsing region moved to the sidebar.workspaces slot); this branch's
sidebar.settings foot seat re-applies on top — the hardcoded foot row
becomes the seat, the seat rides the railIn crossfade, and the shell
spec's renderSlot stub key-splits region vs settings.
Review feedback (tianyicui): 'process' is a poor service name. The family is
now packages/subprocess/ — @deepseek-ai/dsh-subprocess (ctx.subprocess,
abstract SubprocessService, Subprocess* vocabulary) and
@deepseek-ai/dsh-subprocess-local (LocalSubprocessService) — renamed
throughout code, compositions, docs (en+zh, pairs re-recorded), catalogs,
and gates. 'subprocess' is the precise term for managed OS children (the
Python-stdlib sense), avoids colliding with Node's global process object,
and reads as one system beside dsh-subagent-subprocess.
ds-review-bot findings addressed:
- kill() on a settled handle is now a no-op (no signal to a possibly-reused
pgid, no referenced grace timer delaying exit); pinned by a spy test.
- The moved DshEnvironmentKey/DshEnvironment/CollectedOutput types get
drift-checked type-equiv blocks on the new subprocess.md page, restoring
their manifest registration.
- subprocess.md is registered in the core.md sub-page index (en+zh).
The shell is now a pure composition face: no dictionaries, no locale
dependency, and three new chrome content seats (settings.trigger /
settings.header / settings.close) whose slot content also carries the
accessible names (trigger text, dialog aria-labelledby, visually hidden
close label). ui-settings-general returns as the owner of copy that
belongs to no single feature: chrome content, the General section with
its item slot, and the settings dictionaries. Slot types split homes —
trigger/header/close/section live in the shell contract; the
settings.general.item entry moves to the locale package (the common
dependency of every item registrant), with ui-theme consuming it
through a re-export seam; the verbatim duplicate merges are gone and
the dependency graph is a clean DAG.
One highlighter for the client: a synchronous fine-grained shiki core
(JS regex engine, no WASM) in ui-primitives with an explicit grammar
allowlist (typescript, shellscript, json — aliases resolve, unknown
languages take a geometry-identical plain arm). The shared CodeBlock
component owns both arms; markdown fences, the run_code expanded
program body (typescript), and the details panel Input (json) all
route through it. Token colors live in a new ui-theme shiki.css sheet
as --shiki-* custom properties (light/dark blocks), wired through the
shell's base.css chain — tokens-only styling holds; shiki's generated
span tree is the sanctioned innerHTML path (static output, no user
HTML). jsdom specs pin token spans, aliases, both fallbacks, and the
fence route; the built-bundle snapshot asserts the highlighted program
under the code row.
- Mount LocalProcessManager in the sandbox e2e compositions (bwrap/landlock/
seatbelt + the spine multi-project e2e) and add the spine demo's
dsh-process-local devDependency, so SandboxBashExecutor's new inject
resolves when those suites are enabled.
- Extend the Windows test/coverage skip to packages/process/* — the POSIX
process-group suite moved there from packages/bash.
- Update the stale disposal contract: the bash seam JSDoc, BashProcess JSDoc,
and core bash doc (en+zh) now state that composition teardown (the process
manager's disposal) owns kill-and-await, and an executor-only reload leaves
background processes running.
- Record ctx.processes in the architecture capability table and extension map
(en+zh) and the root AGENTS.md layout tree; reword the timeout-library note
so it describes where the plumbing and classification live today.
New process/ capability family: @deepseek-ai/dsh-process owns ctx.processes —
abstract ProcessManager.spawn(spec) over a fully-explicit ProcessSpawnSpec —
plus the shared DSH_* managed-environment and CollectedOutput vocabulary;
@deepseek-ai/dsh-process-local carries the former bash-local run.ts plumbing
(detached groups, tail-keep spill-backed output, credential scrub, kill
escalation, kill-and-join disposal) with no config of its own.
dsh-bash-local becomes a consumer: it keeps command defaulting, the fused
deadline timedOut/aborted classification, the model-friendly terminal env
(now merged through the ordinary env channel), and the [stderr]-marked
background read merge, and spawns through ctx.processes. Background-process
lifetime moves to the manager, so an executor reload no longer kills live
background work; a background spawn failure is injected once into the read
path instead of being buffered as fake stderr. dsh-bash re-exports the moved
vocabulary so bash consumers keep one import root; dsh-bash-sandbox only
redeclares the inherited inject.
Every composition loading a bash executor now loads dsh-process-local (CLI,
examples, python bundled runtime, create-sdk bash feature, inline test
configs).
The tasks/ family now matches the capability-seam shape: @deepseek-ai/dsh-tasks
keeps the abstract TaskService (ctx.tasks contract, vocabulary types, snapshot
invariant companion) and the new @deepseek-ai/dsh-tasks-local carries the
process-local registry (LocalTaskService: in-memory store, settlement,
owner-cleanup effects, teardown, TASK_WAIT_TIMEOUT). Compositions and test
harnesses now load dsh-tasks-local; producers, TaskKindMap merges, and
dsh-tool-tasks keep importing the seam only.
Producer misconfiguration diagnostics name dsh-tasks-local because loading the
implementation is the fix. The registry behavior suite moves to tasks-local;
the seam keeps a stub-subclass registration test and the probe-based invariant
suite.
The web tree had no transient-failure recovery around the loop's model
calls; the TUI agent-spine composition already mounts llm-retry. Same
defaults (2 retries, 500ms->10s backoff). The browser e2e retry scenario
drives it end-to-end: an injected SERVER throw at call 0 recovers through
the durable llm/retry record and completes in the transcript.
Tianyi Cui
kingwl
imccyu