The dev bundle watch missed rebuilds that landed while the registry was
constructing: fs.watchFile captures its comparison baseline with an
ASYNCHRONOUS first stat, so a write racing that window is absorbed into
the baseline and never reported. Standalone repro missed 24/400 same-tick
rewrites; the CI flake in web-plugins.spec.ts ('watch mode: a bundle
content change re-hashes the row...') was exactly this — the spec writes
immediately after createHostWebPluginRegistry returns.
The watch now polls from one registry-owned setInterval against a stat
baseline the scan itself captures synchronously, stat-before-read: a
write landing between stat and read leaves the hash newer than the
baseline (next tick re-hashes to the same rev, no spurious notify); a
write landing after the read leaves the baseline older (next tick
detects and notifies). No blind window. The poll iterates the live
table, so rescans retarget the watch for free and dispose clears one
timer. Stress: real-registry same-tick rewrite 0/600 missed (was 1/300);
spec watch tests 0/50.
New regression test pins the same-tick-as-construction write. Its
rewrite deliberately differs in size from the seed: a same-millisecond
same-size rewrite is invisible to any mtime+size poll (coarse fs
timestamps) — a stat-polling limit, not this regression. Loading-model
Agent Note updated in both languages (pair re-recorded).
The pool selector is defense-in-depth only — pull_request executes the
PR's own workflow definition, so YAML cannot enforce runner trust. Make
the actual enforcement boundary explicit in the decision record:
org-side disabled forking (the public release is an isolated read-only
mirror under a separate org), with migration to a repo-restricted
org-level runner group with base-branch workflow pinning as a hard
gate before forking could ever be enabled.
Keep inherited child prompt markers bounded by the normal silence fallback. Stage web-plugin rescans atomically and retain missing watch state until a successful rebuild.
Sweep all remaining sources that still described coverage as an
enterprise 32-core job: the ci.yml jobs preamble, the three-job
paragraph of the larger-hosted-runners note, and the required-pool
sentence of the portable-recovery note — English and Chinese sides of
both notes, with their i18n pairing records re-recorded.
master gave 2026-06-29-todo-write-tool a Chinese counterpart; the English
side's web-consumer sentences now translate across (in-body links keep
their .md targets per the pairing contract) and the pair is re-recorded.
The todo-panel fake gains the time/callTime fields ToolResultNode now
requires.
- Route untrusted PRs (forks + Dependabot, same author test as e2e.yml)
back to the hosted enterprise pool via a runs-on expression: Dependabot
PRs are same-repo, so the previous head.repo guard admitted
dependency-supplied code onto the persistent self-hosted VM. A single
job with pool selection keeps all-checks-passed free of skips.
- Drop the pnpm-store cache restore from this lane: on self-hosted the
hosted-path cache actually HIT (Linux key) and spent ~52 s pulling
181 MB into a path pnpm never reads; the persistent local store
already serves warm installs in seconds.
- Update the larger-hosted-runners Agent Note (en/zh + i18n pairing
record) so the decision record describes the shipped topology:
coverage on the in-house vm-backup pool for trusted PRs, hosted
Ubuntu 24.04 32-core retained for untrusted PRs.
The projection sentence prescribed resetting todos on window rebuild —
the implementation deliberately preserves the tail-page seed and lets
only in-window/live writes overwrite. The toolview section named a
nonexistent ctx.toolviews registry — the shipped seam is the keyed
conversation.chat.toolview slot via ctx.slots.register. Both sides of
the bilingual pair re-recorded.
The client's todos projection derived only from the paged display window,
so reopening a session whose last todo/write preceded the tail page showed
an empty plan until the user paged back — session-level state cannot be
reconstructed from an arbitrary window. The host owns the full log, so the
tail history response now attaches todos (latest todo/write backscan, the
same posture as the view pairing); installWindow seeds it, window rebuilds
preserve it, and any in-window or live write keeps overwriting it. The
fixture mirrors the host; docs and both Agent Notes record the mechanism.
runtime README documents ConversationSnapshot.todos and its window-scoped
limitation; the todo tool README and Agent Note name the web client among
the event consumers; the web display note records the cold-load gap and
fix directions (bilingual pair re-recorded).
- gen-doc-graphs still summarized the ACP demo as 'text-only'; say
baseline-prompt and regenerate composition.md.
- The llm-catalog note claimed unit coverage for per-agent target
isolation that no test pins directly; state the actual basis
(agent-scoped listener installation) in both languages and re-record.
The shared test module was named harness.ts inside a repo whose product
IS a harness — hopelessly ambiguous. Renamed to scaffold.ts with
launchWebScaffold/WebScaffold; tsconfig plane-split entries, the seam
JSDoc/README mentions, and both Agent Note languages updated.
Both scenarios gain a Playwright interaction step over the settled
transcript (after the golden capture, so committed aria surfaces stay
untouched): replay-round-trip clicks the reasoning fold open/closed over
wire-delivered state; seeded-history expands a read tool row rebuilt from
the cold log and asserts the recorded result text appears (read rows are
expand-in-place — rowExpands routes the click to the inline fold, not the
details column). test:web 30 passed | 1 skipped.
The automation-only rewrite edited many implemented Agent Notes; several
edits replaced still-live or historical rationale instead of reframing:
- llm-model-catalog: restore the prompt/request consistency section and
selection-ownership alternatives — installAgentLlmTarget and the TUI
/model selector still ship that design; only the ACP wire is gone.
- plan-specific-collaboration-state, acp-multi-session, todo-write,
ask-user-question: link the superseding automation-only note instead
of silently rewriting the original decision or motivation; drop a
paragraph duplicating the Web-provider facts stated two paragraphs up.
- sandbox: stop claiming unit coverage for turn-enclosed config writes
(that mechanism left with the bridge) and retitle the commit-boundary
paragraph accordingly.
- Fix the missing blank line before '## Consequences' in the
plugin-command-registration pair, the JSON-RPC/Web render-intent
consumer misattribution (the second consumer is the host/client
runtime), stale bash_output/bash_kill names, and 'optional goals' in
architecture.md.
- examples/acp-agent/README.md: point at the package contract instead
of restating it; packages/ui/permission and plan-mode READMEs record
the consumer-less preset service and the exit_plan_mode coverage gap
under Known Limitations.
- 2026-06-19-acp-snapshot-tests: the new note defers the corpus
migration rather than committing to it; say so.
Re-record the touched bilingual pairs.
ACP v1 requires every agent to accept text AND resource_link prompt
content; the automation rewrite dropped the resource_link half of that
baseline. Restore the old bracketed-reference flattening in the codec,
reject only beyond-baseline blocks, and update the package contract and
Agent Note.
Also release the per-session prompt slot when agent.send() throws
synchronously (an agent disposed outside the bridge would otherwise
wedge the session into permanent 'already in flight' rejections),
drop the tautological version-negotiation branch, prove the scenario
env layer reaches the snapshot subprocess, pin bridge-side fail-closed
permission errors, and correct two overpromising test names.
Chinese counterpart translated per the terminology table and the
2026-07-18 TUI note's register; switcher lines added on both sides; pair
recorded. doc-sync 24/24.
Simplify the Commander adapter now that behavior can change: dispatch a leading
`web` token to its own parser instead of a subcommand of the root program, and
read opts()/processedArgs after parse() instead of action closures with a
mutable holder.
This removes enablePositionalOptions(), the parent-option leak guard, both
action closures, and the --resume/--prompt argParser threading. Behavior
changes: `dsh -p x web` is a headless prompt (extra positional dropped),
`dsh web -p x` fails loud (web has no -p), and a repeated --resume is natural
last-wins. The two real fail-loud invariants stay as post-parse checks: an empty
--resume= id (agent-loop treats '' as no-resume) and an empty -p task.
Trims args.spec.ts to the routing/fail-loud/help behavior that matters; the
tui-agent keyless PTY smoke still covers bin.ts dispatch end to end. Net ~114
fewer lines across adapter and tests.
testing.md gains the web browser snapshot tier entry (divergent
DSH_SNAPSHOT=... test:web commands) and names apps/web/tests/snapshots/
as the web surface's snapshot home. The GUI testing note's tier map and
lane map gain the e2e scenarios (both languages, pair re-recorded) and
drop the stale verify-session-real references (those scripts left with
the missions/ tree). packages/client/AGENTS.md check ladder covers the
wire-carriage trigger and refresh/record commands. acp-snapshot README
stops claiming the whole package is ACP-specific — its normalizers are
transport-neutral with three consumers now. vitest.web.config.ts header
carries TODO(ci-browser) with the staged-reversal pointer.
The design-study Agent Note moves proposed/ -> implemented/ rewritten in
present tense: all review decisions recorded (llm:false seam over the
placeholder-key hack, providers-mode replay, whenIdle barrier stack,
single aria golden + anchors, TUI-style inline modes over a suite
factory, scrub-only header stance, CI deferral) with re-entry triggers
under Deferred.
Address ds-review-bot: the bounded-request-recovery Agent Note stated the
shipped default carried four transient codes, and the llm-streaming contract
omitted the new cross-adapter empty-response classification. Update both
current-state contract docs to the five-code default and cross-link the
empty-response bug-fix note.
Replace the dsh CLI's three hand-rolled parsing idioms (raw argv[0]/includes
dispatch in bin.ts, per-mode node:util parseArgs in headless.ts/web.ts, and the
bespoke parseResumeArg scanner in dsh-app-boot) with a single Commander adapter
in apps/cli/src/args.ts. parseDshArgs resolves argv into a discriminated
DshInvocation union; bin.ts switches on the mode and dynamic-imports the chosen
module, which now consumes already-parsed values.
- web is a real subcommand; --host uses choices and --port an argParser range
check, moving validation into the parser.
- --resume rejects empty and repeated forms; --prompt rejects empty; a config
positional after --prompt and a root flag placed before web fail loud.
- adds --help/--version; removes parseResumeArg from dsh-app-boot.
- new apps/cli/tests/args.spec.ts (apps/*/tests added to vitest include,
apps/cli/tests to tsconfig.host.json); the tui-agent keyless PTY smoke covers
bin.ts dispatch end to end unchanged.
A well-formed provider stream that ends with finish_reason stop and zero
content blocks previously became a successful empty assistant message: the
turn completed silently, and drivers like goal-session counted the no-op
round. Both adapters now map that degenerate completion to a finish
{kind:'error'} with the new canonical EMPTY_RESPONSE code from dsh-llm, and
dsh-llm-retry adds the code to its default retryable set, so the existing
closed-step recovery path retries it and fails loud once the budget is
exhausted.
Covered by adapter unit tests, an llm-retry default-policy test, and a new
authored keyless ACP snapshot (empty-response-retry) with a deterministic
1 ms zero-jitter retry overlay.
Design study for a deterministic, keyless browser e2e lane over the real
assembled web chain (chromium -> SSE/HTTP wire -> apiproxy -> agent loop ->
persistence), replayed through dsh-llm-replay from recorded session-log
fixtures, with aria-tree goldens plus in-process world-state assertions.
Synthesized from an OSS prior-art survey (LibreChat, ai-chatbot, lobe-chat,
OpenHands, cline, aimock...), a repo seam deep-dive, and three adversarial
critiques (doctrine, flakiness, YAGNI). Records the settled shape (no new
package, no suite factory, seed via the real persistence API, whenIdle
barrier stack, no transient-DOM assertions) and the open questions (LLM
seam, Loader-izing dsh web, header pin, golden breadth, settled signal).
