master introduced AttachmentStore.saveImages as the batch admission
(count/aggregate-byte/media-type limits, validate-all-before-save,
ordered commit). admitEncodedImages narrows to the shared wire entry:
canonical-base64 enforcement plus delegation to saveImages, keeping one
home for batch policy while both wire endpoints (prompt RPC and the
command executor) still call one function. Test doubles gain saveImages;
batch-limit error texts follow saveImages' wording.
A claimed slash command consumed only the text half of the composer
submission: /goal with reference images executed, cleared the draft, and
silently stranded the images in the rail. Model-visible attachment intent
had no route through the command plane.
The submission envelope is now modeled end to end. CommandDefinition
input.images declares acceptance; the declaration rides the descriptor to
every client, onto the minted CommandClaim, and into the input machine's
claim snapshot. commands.execute carries the submission's base64 images
and enforces the declaration in the executor: non-declaring commands, a
missing attachment store, and exceeded batch limits settle as logged
error results before the handler runs. Admission reuses the attachment
package's new admitEncodedImages, extracted from api-proxy's prompt path
so both wire endpoints share one limits/validation/commit sequence.
Producers own model visibility: /goal submits one user followup (image
blocks + a fixed reference line) after a successful create/edit so goal
rounds read the images from session history; /plan folds them into its
steered message. Grammar misfits (/goal pause, bare /plan, /plan off)
return direct errors and the composer keeps the images.
On the client, enter adjudication carries a SubmitEnvelope and every
command route that cannot consume images throws a localized refusal that
renders as one composer notice with draft and images retained; the
claimed pre-gate applies the same copy. An accepting claim serializes the
draft images, forwards them to commands.execute, and clears plus releases
them only on a success outcome.
The assembled web test roster gains the ui-input-trigger and ui-commands
plugins, mirroring the shipped composition, so slash submissions exercise
the command plane; a new keyless snapshot pins the refusal banner and the
accepting /goal flow over the built client graph.
Replace the legacy reference.* API Proxy domain with @Remote methods on the
owning services, following the typert gateway design master adopted on
2026-08-02 (message-feedback and plugin-inventory precedents):
- FileReferenceService and SessionReferenceResolver extend TypertRemoteService;
fileReferences/list and sessionReferenceResolver/candidates are unary Remote
methods cancelled through the reserved trailing signal, and the candidates
face attaches each candidate's canonical mention under the configured limit
- move the wire types to type-only ./types subpaths (FileReferenceCandidate,
SessionReferenceMentionCandidate) and export ./typert plus ./remote artifacts
- mount both contributions in the api-remotes client assembly; ui-reference
consumes ctx.remote instead of connection.api.references and registers zh/en
locale dictionaries for its sections and labels
- delete the reference.* routes, schemas, map rows, client stubs, and fixtures;
the connection fixture serves the Remote endpoints instead
- release deliverPrompt admission listeners when the agent is disposed with the
prepared prompt still pending, and cover the reference-* RpcError codes in
the schema spec
- add the missing tsconfig paths for the /grammar and /types subpaths (clean-
tree vitest could not resolve @deepseek-ai/dsh-file-reference/grammar)
- regenerate the cordis catalog, capability seams, and event matrix; update the
owning bilingual READMEs, Agent Notes, and the reference-composer golden
A max-tokens response that included a tool call persisted assembler-transformed
content next to replay metadata projected from the untransformed native message,
so the next request died in history reconstruction with INVALID_REPLAY_STATE and
the session stayed permanently stuck.
Write side: the finish chunk's replayState becomes a typed ReplayEnvelope —
opaque response-level metadata plus optional per-block entries aligned with the
emitted block sequence. BlockAssembler computes one keep/drop decision for
blocks and entries together, so stored metadata always describes stored content
and retained blocks keep their signatures. pi-ai splits its state into a
version-2 response half and per-block signature entries.
Read side: durable content is authoritative. toPiAssistant degrades any
unusable state — foreign kind, other versions (including the flat v1 form
already on disk), malformed metadata, or content/block mismatches — to the
existing provider-neutral conversion with an onReplayDegrade diagnostic instead
of failing the request, which un-bricks sessions poisoned before this change.
Covered by assembler and replay unit tests, an agent-loop continuation
regression, keyless real-composition continuation tests (native pruned-envelope
replay and legacy flat-state degrade), and the authored keyless snapshot
scenario max-tokens-continue through the assembled ACP app.
The shipped bundles keep ctx.sessionQuery mounted but set the new
session-query-sqlite `openAt: never` phase: searchSessions/searchEvents
fail with the typed SESSION_QUERY_SEARCH_DISABLED code before any request
normalization, node:sqlite is never imported or opened, and no source
observation or reconciliation runs. Every inherited exact read, filter,
and trace — session export descendants, subagent-fork Workspace
inheritance, title reads — keeps working, and the Web sidebar search
degrades to its designed local title/workspace matching. Enabling content
search is a one-line openAt override in a later patch layer; the web e2e
scaffold keeps it enabled as the assembled opt-in coverage.
turn/start precedes the first claim and pre-step, a rejected or empty first claim closes a zero-step turn, llm/stream and tools/* are their own domains, and agent/turn-stopping is serial. Regenerates the lifecycle diagram from the same facts and names --patch overlays exactly.
Correct the reworked English wording in the CLI profile section, architecture map, and llm-streaming cross-reference, then bring every edited pair's Chinese side along and re-record the pairing hashes.
Keep both Remote contributions master and this branch add: the mount loop
now carries commandsRemote, goalsRemote, pluginInventoryRemote, and
messageFeedbackRemote, with both new tsconfig references retained.
Overlaps master on the preset e2e, the web scaffold, and the generated
config-catalog triplet; all merged textually, so the catalog is regenerated and
its pairing hashes re-recorded rather than trusted.
Adapt to two contract changes master introduced:
- The generated Remote face now wraps every business result in
RemoteResult, folding carrier failures into an ok:false branch instead
of rejecting. The controller reads that envelope at its three call
sites and maps a carrier failure onto the same settled shape the
controls already render; three specs cover the new branch.
- Client packages split their tsconfig into host and client halves, and
the host aggregate now compiles any test not named *.client.spec.*.
Rename this package's specs to the client convention and drop the
../connection project reference, which pointed at a solution file that
no longer carries the client sources.
Keep master's mount loop with its rollback-on-failure in api-remotes and
add messageFeedbackRemote to it.
Resolve the note-erasure race the review found: a control that rendered
before the first list read held no item and passed note=undefined, so
switching a rating silently dropped the stored note. The controller now
owns note resolution and toggle-vs-retract, deciding from the committed
item inside the serialized mutation, and clearNote expresses deletion.
Also from review:
- serialize the reconnect re-read behind queued mutations (resync), so a
list reply cannot resurrect a version a newer mutation replaced
- re-check disposal after the seeding read, so an unloaded fiber never
reaches the wire
- drop Object.freeze on Maps, which does not prevent set/delete
- declare the @deepseek-ai/dsh-client-connection dependency it imports
- surface a failed list load in the controls
- drop the /client value exports that had no consumer
- state the per-turn render scope in the README and subsystem pages
- align the package version with the root
Adding COLLAPSE_SECTION_ORDER and CODE_ONLY_INSTRUCTION shifted every
later line in packages/core/tools/src/index.ts by 14, and three generated
artifacts cite those lines: the subsystems cordis-surface region, the
event producer/consumer matrix, and the Agent Note pair record left
inconsistent by the cherry-pick resolution.
wireSchemas() already advertised only run_code under mode: 'code', but the
executor resolved every call through get(), which returns the full visible
map plus the reserved transport. A model could name a native tool directly
and bypass run_code entirely. Route the execution-path lookups through a
new private resolveExecution() that applies the mode collapse at the
operation boundary: model-direct calls under 'code' may only name run_code
(UNKNOWN_TOOL otherwise), while SDK sub-dispatches (parent token set) keep
every visible tool. get()/schemas() public semantics are unchanged.
The denial happens at createExecution, before the extensible policy
pipeline — pre-execute listeners, approval ask, and guards never observe
a call that is deterministically denied. A collapsed call honors the
pre-dispatch cancellation contract, routes aborted results through the
visible tool's finalizeContent, and captures the finalizer before
argument materialization.
Regenerated docs, catalogs, graphs, scoped events, re-recorded
translation pairs, and updated test assertions.
Fixes#1815
wireSchemas() already advertised only run_code under mode: 'code', but the
executor resolved every call through get(), which returns the full visible
map plus the reserved transport. A model could name a native tool directly
and bypass run_code entirely. Route the execution-path lookups through a
new private resolveExecution() that applies the mode collapse at the
operation boundary: model-direct calls under 'code' may only name run_code
(UNKNOWN_TOOL otherwise), while SDK sub-dispatches (parent token set) keep
every visible tool. get()/schemas() public semantics are unchanged.
The denial happens at createExecution, before the extensible policy
pipeline — pre-execute listeners, approval ask, and guards never observe
a call that is deterministically denied. A collapsed call honors the
pre-dispatch cancellation contract, routes aborted results through the
visible tool's finalizeContent, and captures the finalizer before
argument materialization.
Under code mode, a system-prompt/assemble listener filters out tool:*
guidance sections that told the model to call native tools directly.
The tools:sdk section and SDK types remain so programs can still use
all tools through run_code.
Regenerated docs, catalogs, graphs, scoped events, and re-recorded
translation pairs.
Fixes#1815
