Responding to review on #661: a public method on the generic ToolRegistry
service whose only caller is the run_code bridge was ad-hoc surface
widening. The bridge now receives it as a registry-private capability
closure in RunCodeBridgeOptions (the requireRuntime idiom, alongside the
cap), the method is private, and it leaves the generated service
catalog/API surfaces.
The pattern is now named as a code smell where reviewers look: the
packages/AGENTS.md capability-interface rule gains the inverse-smell
clause (ceiling 660→675 — the list is at capacity and the clause needs
one sentence), and dsh-code-review's capability-fit check tells reviewers
to flag single-consumer public service methods and require the closure
form.
Responding to ds-review-bot round 2 on #661:
- logWork is bounded: past maxParallelSubCalls pending shaped-append tasks
the ordered commit lane holds (Promise.race drains one), so a slow spill
backend backpressures the run instead of accumulating unbounded pending
I/O and retained results. Tasks self-remove on settlement; run
settlement still drains every task inside the open turn. New spill test
drives three oversized reads against a hung backend at cap 1 and proves
the third dispatch cannot start until a save drains.
- shapeDispatchLog's catch uses errorMessage() (total), so a thrown value
with a throwing toString cannot escape the containment and lose the
settle event.
- CodeDispatchLog.content documented as the RENDERED result projection
(native tool/result vocabulary), not what the program received — the
program gets the structured value; doc pair + type-equiv re-synced.
- New RECORDED tui-agent snapshot scenario code-mode-dispatch-spill: the
real Loader-visible composition (worker runtime + spill-local + policy)
drives an oversized bash sub-call end-to-end; replay proves the durable
dispatch copy is bounded to preview + locator while the program value
stays whole (the outer result carries just the line count).
Agent Note updated (both languages).
Conflict resolution: drainDispatches composes both sides — the reworked
single-lane driver's drive() (ordered stages, mid-flight commit drain)
followed by this branch's logWork drain, so every shaped settle append
still lands inside the open run_code turn.
Responding to ds-review-bot round 2 on #658 (three critical findings, one
warning — all rooted in the pump/commit split racing ordered stages):
- ONE driver lane now owns every ordered stage: the start append, prepare
(pre-execute/guards), and the head-of-line commit (post-execute, context
deferral, settle append). start() is awaited before the next entry can
start, so concurrent submissions can no longer run pre-execute pipelines
concurrently; only the around-dispatch/body stage overlaps, matching the
native loop's fillPool sequencing.
- An exclusive call's barrier now holds through its COMMIT: later starts
wait for the exclusive pipeline (post-execute included) to finish, the
native exclusive-group semantics.
- drainDispatches() awaits the driver run itself, so a commit already
mid-flight when the program returns is drained before run_code closes
the turn — the settle event and deferred contexts land inside it.
- maxParallelSubCalls is resolved and validated at construction (positive
integer), so direct construction can no longer wedge the pool with 0.
New tests: overlapping-submission ordered-prepare, barrier-through-commit,
drain-mid-commit, cap rejection. 96 keyless snapshots replay unchanged;
Agent Note updated (both languages).
The coverage gate flagged three untaken paths in the bridge scheduler:
- the exclusive-head inFlight re-check was dead (the shared guard above
already returns for an exclusive head with any in-flight sibling) — removed;
- the commit-cursor undefined-dispatched break was structurally unreachable
once entries join commitQueue only after start() ran synchronously —
reordered the pump so the invariant holds by construction, annotated;
- the finish (final-result) commit arm and the pump re-entry guard gain a
covering test (throwing tools/pre-execute listener) and a defensive
annotation respectively; mid-run unregistration test renamed to match its
actual post-result settlement path.
Also covers the direct-construction maxParallelSubCalls default (index.ts)
and commits the regenerated persistence catalog for the new dispatch pair.
ds-review-bot findings on the parallel bridge: sub-dispatches now run
through the registry's staged scheduler view (the loop's own seam) —
ordered prepare (pre-execute/guards) at submission-ordered start time,
only the dispatch/body stage overlapping, and a head-of-line commit
cursor running post-execute, context deferral, and the settle event in
submission order (new spec pins post order + context order under
out-of-order completion). Queued dispatches reclassify via
executionMode() immediately before starting, so a registry mutation
while queued flips them exclusive (native lazy-reclassification
semantics). Config and tool catalogs regenerated; the tool-catalog
metadata now names the start/settle pair and the scheduling contract.
ds-review-bot findings: the packaged Python runtime smoke's scripted
run_code call gains the required description; the ToolDefinition JSDoc
and the Code Mode foundation note (both languages, pair re-recorded) now
state both required parameters; the cordis-dynamic-toolchain fixture's
request/header line is re-compacted so the header-scrub hygiene guard
passes (my earlier patch had re-spaced it). The TUI terminal fixture was
already regenerated from keyless replay in the previous commit.
New tools/code-dispatch-log waterfall (run via registry.shapeDispatchLog,
contained — a throwing listener falls back to the unshaped content) lets
listeners reshape the tool/code-dispatch event's content before the
bridge appends it. dsh-spill-policy registers a second arm sharing the
model-facing arm's exact replacement pipeline (same maxInlineBytes cap,
preview + locator, within-cap invariant, best-effort fallbacks), with
artifacts labeled dispatch under the sub-call id. The program's value is
untouched; read sub-calls ARE bounded (a log copy is not model context,
and read produces the biggest logs). Resolves the tools README's
uncapped-dispatch-log Known Limitation.
The bridge replaces its serialization queue with a pool that reuses the
native concurrency contract: submissions classify through
registry.executionMode (fail-closed isConcurrencySafe), start strictly in
submission order, overlap up to the validated maxParallelSubCalls config
(default 10; 1 restores serial), and exclusive calls drain the pool, run
alone, and bar later calls. Each started sub-call logs a
tool/code-dispatch-start event at pool entry; the existing
tool/code-dispatch settles the pair (started ⇔ settles exactly once;
abandoned queued calls log neither). SDK prompt guidance now states the
true Promise.all contract — re-recorded across every code/both-mode
snapshot (plus the stale cordis-dynamic-toolchain fixture gaining the
required description arg).
Client: CodeSubCall widens to RunningToolCall | ToolResultNode — starts
land the running shape (rows wear the native running ring), settles
replace in place preserving start order, callTime pairs to the start
time. Fixture emits start/settle pairs; jsdom pins the running sub-row;
runtime specs pin in-place settlement and out-of-order completion.
run_code gains a required bash-style description parameter: presentCall
titles the card with it and moves the program to rawInput, so every
surface gets a readable label. tool/code-dispatch now logs each
sub-call's complete content/isError (the tool/result vocabulary),
replacing the bounded resultSummary and deleting the summarize/cwd
machinery — a UI renders sub-calls through the identical path as native
results. The dsh config tree mounts the worker code runtime and reads
DSH_TOOLS_MODE (temporary seam until per-session mode selection lands).
Session format stays v0 (pre-release churn). Code-mode ACP/TUI fixtures
re-recorded; TUI presenter pin refreshed; catalogs regenerated. Keyless
web smoke pins the code-mode wire contract (tools=[run_code] + SDK
prompt section).
Master advanced 11 commits mid-merge (the Code Mode registry integration:
mode config, run_code, the tools:sdk section, the ACP run_code cards and
unified demos). The fusion makes Code Mode scope-aware end to end:
- wireSchemas(scope): the mode-aware wire contribution is computed from the
CALLING SCOPE's visible set (scoped tools join, shadowing and restrictions
apply) and feeds the {schemas, knownNames} provider protocol.
- knownNames under the mode collapse: a per-scope RESTRICTION is runtime
state, so the universe stays pre-restriction (a restricted-away tool in
toolOrder is a normal absence) — but the MODE collapse is deployment
config, so under mode 'code' the universe is [run_code] and a toolOrder
naming a native tool fails every assembly loud (master's tested decision,
kept).
- The tools:sdk section renders per assembly CONTEXT: the SDK declares
exactly the calling agent's callable set, using the section-text provider
signature this branch already had.
- run_code bindings enumerate schemas(exec.agent) — a program can bind
exactly what its prompt promised; sub-dispatches already threaded
exec.agent through registry.execute, so scoped resolution and carriers
flow unchanged.
- dsh-tools declares both sides' new deps (dsh-scope + dsh-session);
lockfile and all generated catalogs/graphs/api-catalog regenerated.
Gates green on the merged tree: typecheck, lint, per-file 100% coverage
(2710 tests), snapshots (41), doc-sync, module graph, build, hygiene, demo
smoke.
Systematic trace through Zed (crates/agent_ui thread_view.rs +
crates/acp_thread): kind:execute routes a tool call onto the
terminal-card layout, whose header (render_collapsible_command) has NO
disclosure toggle, whose body content renders only when is_open — a
flag only a real terminal entity can ever set — and which suppresses
the Raw Input view outright. Every prior attempt (rawInput, pending
content, completed content) targeted slots that layout structurally
never renders; the one slot it always shows is the TITLE, which said
"Run code". codex-acp confirms the idiom: execute cards are titled
with the command itself.
presentCall now titles the card with the program (rawInput kept as the
canonical input slot); presentResult omits the title — an update
replaces only provided fields, so the program header persists — and
carries the captured output as content. Goldens re-recorded; the unit
test pins title-carries-program on both frames.
The previous commit put the fenced program only on the pending card —
but an ACP tool_call_update's content REPLACES the card content (Zed
truncates to the new list, crates/acp_thread update_fields), so the
code vanished the moment the run completed and was effectively never
visible. presentResult now re-carries the fenced program before the
captured output via a shared fencedProgram helper; the completed card
body is program + output, rendered by Zed as syntax-highlighted
markdown behind the card disclosure. Goldens re-recorded (filtered
this time: DSH_SNAPSHOT=record vitest -u -t mode-turn); unit test pins
the two-block result content.
The generated program rode only rawInput — the detail/expanded slot
many ACP clients never open — so the code a run executed was invisible
in the UI stream. presentCall now also carries it as a fenced ts block
in the card's content, which the bridge already forwards as tool_call
content. The two code-mode snapshot goldens are re-recorded live and
replay green; the presentation unit test pins the fenced block.
[P1] review finding: the run-scoped abort + queue drain ran only after
runtime.run() FULFILLED, so a backend that starts a binding call and
then throws left the sub-dispatch running past run_code's settlement —
its tool/code-dispatch event could append after the parent call
returned, breaking the drain-before-return contract. The quiescence
pair now lives in a finally around runtime.run(); the folded queue tail
keeps the drain from masking the thrown error. Pinned by a test whose
fake runtime fails mid-flight: pre-fix it returns in milliseconds with
the slow tool still running.
Both bot criticals verified against the code and rejected as exploit
paths — pinned instead of patched:
The bindings loop already excludes run_code (the skip predates the
finding), and the runtime host resolves forged port calls as own
properties of the bindings record, so an absent binding is unreachable
from a program under any mode. A new both-mode test pins the invariant:
the record has no run_code key on any lookup path.
The drain await cannot mask a run failure: `queue` is the folded tail
(every link swallows its rejection), so `await queue` never rejects and
the runtime's own result.error always reaches the CodeRunFailedError
conversion — the existing abort test exercises exactly the
queued-abandonment-plus-run-failure scenario. Stated at the drain site so
the fold's purpose is explicit.
Tianyi Cui
Yichen Jiang