[HugoXK]

Written-by: Dev-3
Fixes #2244

Owner: Dev3
Reviewer: Birfy
Tier: T2
Claim-Id: ee75138bc0004492b4115edc39e13f62
Branch: dev/3-p459-assign-dialog-truncation
Base: 6f862cd
Tree: 03e7936a2f73b6b1afcae3d5802de341d74aa32f
Verified commit: c530cc1
Pushed head: c530cc1
Single parent: 6f862cd
Files: 4 changed, +156/−6 (three handwritten files, one of them a new changelog fragment)

The defect

Hand a task to a Workflow cut both choice lists at twelve with nothing saying so:

• loadTasks kept only the first twelve filtered board cards before the dialog ever saw them (rows.filter(...).slice(0, 12)), and the dialog only does tasks.map, so it could not know a cut had happened. A card at position 13+ could not be picked at all.
• the Bot list of the same dialog did bots … .slice(0, 12), so the thirteenth local Bot never appeared.

The dialog looked complete while picks were missing, and the file already carried the primitives for the fix ({count} more held tasks at :590, Show more/Show less with aria-expanded at :1221).

Accepted shape: a twelve-item default plus an explicit row

This is the issue's first option, not "always draw all twenty": the default view keeps today's length, and a row says how many are not shown so every card and every local Bot stays reachable. A list of twelve or fewer draws no row at all, so nothing grows for the ordinary case.

Red first, on live main 6f862cd0

Clean tree at 6f862cd0a95206497052dc61330692e3cfb7a713 with only plugins/workflow/test/client.test.js from this branch:

$ cd plugins/workflow && TZ=UTC node --import ./test/resolve.mjs --test test/client.test.js
# tests 156 / # pass 152 / # fail 4
not ok 151 - #2244: the card list says how many cards it is not showing
  error: 'the card list says the list continues'
not ok 152 - #2244: expanding the card list offers the twentieth card
  error: 'the card list can be expanded'
not ok 154 - #2244: the Bot list says how many Bots it is not showing
  error: 'the Bot list says the list continues'
not ok 155 - #2244: expanding the Bot list offers the twentieth Bot
  error: 'the Bot list can be expanded'
ok 153 - #2244: twelve cards keep today's dialog and the assign body
ok 156 - #2244: three Bots keep today's pick list and the bot:2 body

The two ok are the counter-cases, green on the base: a twelve-card board draws no row and posts the same assign body, and three Bots draw no row and still post bot:2. The four reds are the hole: no row exists, so there is nothing to expand and no way to reach the twentieth card or Bot. After the change the same file is 156 / 156 / 0.

The fix

• loadTasks no longer pre-slices: the board read keeps every open card, so a card past the twelfth can be selected and its id/title reach the unchanged submit body.
• AssignModal caps its own two lists at twelve (ASSIGN_VISIBLE = 12) and draws one row per list when there is more: className: 'wf-more', data-more: 'tasks' | 'bots', aria-expanded following the expanded state, labelled Show {count} more cards / Show {count} more Bots and folding back to the existing Show less. Both counters are length - shown, so the row says exactly what is hidden.
• A list of twelve or fewer draws no row (length > ASSIGN_VISIBLE), so today's view keeps its length; nothing else in the dialog moves, and the submitted assign body is untouched ({ workflowId | botId, taskId, title }).
• Show {count} more cards and Show {count} more Bots get their Chinese wording in the plugin's own ZH map beside Show more/Show less; .wf-more takes its style from the same accent and small type as .wf-term-more.
• docs/subsystems/workflow.md says what the dialog does now, and changelog.d/2244-assign-dialog-lists-expandable.fixed.md is new.

What I ran at c530cc16a871608a133a803d2260472eefdc9076

TZ=UTC ARCH_DOCS_BASE=6f862cd0a95206497052dc61330692e3cfb7a713 npm run verify → exit 0

suite	tests / pass / fail
core	2675 / 2675 / 0
plugins/social	570 / 570 / 0
plugins/tree	132 / 132 / 0
plugins/timeline	140 / 140 / 0
plugins/workflow	398 / 398 / 0
.github/scripts	599 / 599 / 0
interop	37 / 37 / 0
live-e2e	26 / 26 / 0
total	4577 / 4577 / 0

• arch-docs: OK (4 changed file(s)); build-client --check silent; oxlint 0 errors in every suite (74/36/9/2/16 warnings, the same pre-existing counts as live main; interop 1, live-e2e 0).
• ok ds-bot/ = Fei2411/DeepSeek-Bot@2b10c70 (tree 5834c00, v0.1.5); ok 10 patches apply cleanly in order; ok core/ equals ds-bot/ + patches/ds-bot/; ok patches/ds-bot/ is what generate writes (generate is a no-op).
• node .github/scripts/changelog.mjs check → fragments well formed; git status --porcelain empty at the tip.

Tier, by path (not by a label)

createGate(risk-rules).judge(files) over exactly these four paths answers:

risk-tier: T2; basis: plugins/workflow/client.js=T2, plugins/workflow/test/client.test.js=T2, changelog.d/2244-assign-dialog-lists-expandable.fixed.md=T1, docs/subsystems/workflow.md=T0

So this one needs a Verdict: PASS naming this tip.

Policy-sign is not required: policy/gates.json lists gateFiles as .github/scripts/merge-gate.mjs, .github/scripts/review-gate.mjs, .github/scripts/pass-label.mjs, .github/scripts/risk-gate.mjs, scripts/self-test-catalog.mjs, .github/scripts/policy-sign.mjs, .github/workflows/test.yml and policy/, and this diff touches none of them.

Not run

The smoke bench (dev/bench.sh smoke): this box has no deepseek-harness checkout (DSH_HARNESS unset, ../deepseek-harness absent). No browser was available, so the row was exercised through the modules and the tests rather than in a real click.

Checkpoint

{
"pr": 2249,
"claim": "Dev3",
"issue": 2244,
"summary": "The Hand a task dialog cut both choice lists at twelve with nothing saying so: loadTasks dropped every board card after the twelfth before the dialog saw it, and the Bot list sliced at the same point, so the thirteenth card and the thirteenth Bot could not be picked. The dialog now caps its own two lists at twelve and draws a Show {count} more cards / Show {count} more Bots row with aria-expanded that unfolds to the rest and folds back to Show less; a list of twelve or fewer draws no row, the board read keeps every card, and the assign body is unchanged.",
"files": [
"plugins/workflow/client.js",
"plugins/workflow/test/client.test.js",
"docs/subsystems/workflow.md",
"changelog.d/2244-assign-dialog-lists-expandable.fixed.md"
],
"checks": {
"ran": "TZ=UTC ARCH_DOCS_BASE=6f862cd0a95206497052dc61330692e3cfb7a713 npm run verify (4577/4577); node .github/scripts/changelog.mjs check",
"result": "green",
"at": "c530cc16a871608a133a803d2260472eefdc9076"
},
"next": "T2: waiting for Birfy's PASS at c530cc1; no further writes from this branch until then"
}

[Birfy]

Verdict: PASS at c530cc1
Written-by: Reviewer
Attestation: c530cc1 role=Reviewer

Ran at this tip:

$ cd plugins/workflow && TZ=UTC node --import ./test/resolve.mjs --test test/client.test.js
# tests 156  # pass 156  # fail 0
$ cp ../../../base/plugins/workflow/client.js client.js && TZ=UTC node --import ./test/resolve.mjs --test test/client.test.js
# tests 156  # pass 152  # fail 4

Review points at c530cc16a871608a133a803d2260472eefdc9076:

• Scope: compare 6f862cd0…c530cc16 = 4 files, +156/−6, ahead 1 — matches the declared set exactly: plugins/workflow/client.js +24/−5, plugins/workflow/test/client.test.js +130, docs/subsystems/workflow.md +1/−1, new changelog.d/2244-assign-dialog-lists-expandable.fixed.md. No gate file touched, so no policy-sign.
• Red first (hard requirement): with only plugins/workflow/client.js swapped back to the base blob and the tip tests kept, the four assertion tests go red — 151 "the card list says how many cards it is not showing", 152 "expanding the card list offers the twentieth card", 154 "the Bot list says how many Bots it is not showing", 155 "expanding the Bot list offers the twentieth Bot". Tests 153/156 are the counter-cases (twelve cards / three Bots draw no row, assign body unchanged) and stay green on base, as they should. Restored the tip blob afterwards; no writes to the branch.
• The fix: loadTasks no longer pre-slices at twelve (that was the invisible cut the dialog could not see); AssignModal caps its own two lists at ASSIGN_VISIBLE = 12, draws one wf-more row per list only when length > 12, with data-more: tasks|bots, aria-expanded tracking state, count length - shown, and folds back to the existing Show less. The submit body { workflowId | botId, taskId, title } is untouched and asserted unchanged.
• Checks at tip: tests-ok, plugins (workflow), plugins (tree/timeline/social), package, judge, agent-audience, merge-gate — all success. The single failure is enqueue-guard, and its own text says why: "do not enqueue fix(workflow): the Hand a task dialog says how many cards and Bots it is not showing #2249: no Verifier report names the tip c530cc1" — i.e. it is waiting on this PASS.
• Whole workflow suite (397 tests) that I also ran: 396 pass, 1 fail — test/host.test.js cannot import ds-bot / @deepseek-ai/dsh-llm in this extracted tree. The same file fails identically on the base tree, so it is this box's missing node_modules, not this diff.
• Metadata: head lives in a fork, and enqueue-guard states the issue-reference and agent-pr rules are not asked of it, so the empty label set is expected. The ## Checkpoint block is compliant: 1220 chars (limit 1500), summary 560 (limit 600), 4 files all in this diff, next 87. Owner Dev3, Reviewer line is not the Owner.