fix(wizard): no cancelar ni generar ante preguntas de stop o de etiqueta

[cursor]

Resumen ejecutivo

Qué cambia

El Wizard ya no trata como orden una pregunta educativa sobre cancelar/parar ni una pregunta sobre la etiqueta «generate video». Esas frases vuelven a ser how-to: no cancelan el job activo y no encolan GPU.

Para qué sirve

Tras #279, «How do I launch the video?» ya no genera, pero «How do I stop video generation if it fails?» seguía emitiendo cancel_task, y «What does the generate video button do?» seguía emitiendo prepare_video + start_generation si había prompt en Studio.

Impacto para el usuario

• Preguntar cómo o cuándo parar/cancelar ya no mata la generación en curso.
• Preguntar qué hace el botón Generate ya no lanza un vídeo.
• «cancela el vídeo», «Can you cancel the generation?» y «Generate a video of a wizard» siguen ejecutándose.

Riesgo

• Bajo
• Medio
• Alto

Estado

• En desarrollo
• Listo para revisión
• Bloqueado por CI o revisión
• Requiere migración o acción manual

Summary

isHowToGenerateQuestion only covered generate/create/launch/start/make. English stop/cancel and Spanish cancelar still matched inside questions, and reconcileAgentTurnWithRequest forced cancel_task with confirm: true, which executeAgentActions sends to the canonical queue.

UI-label questions («What does the generate video button do?», «¿Qué hace el botón genera el video?») still matched EXPLICIT_VIDEO_REQUESTS / isResumePreparedStudioVideoRequest because they contain generate/genera + video.

This PR extends the how-to window with stop/cancel questions and UI-label generate questions, and isExplicitCancelRequest now rejects how-to text. Imperative cancel and generate commands are unchanged.

Overview

Wizard text classification only. No backend, receipts, persistence, or Studio form changes.

1. User has a running GPU job and asks «How do I stop video generation if it fails?» → read-only reply, no cancel_task.
2. Studio has a prepared prompt and the user asks «What does the generate video button do?» → no start_generation.
3. «cancela el vídeo» / «Generate a video of a wizard» still reconcile to the executable actions.

Detailed changes

Backend

Unchanged.

UI and Wizard

• agentActions.ts: how-to patterns for stop/cancel questions and generate-button labels; isExplicitCancelRequest returns false when how-to matches.
• agentActions.test.mjs: locks both directions (question vs command).

Data, provenance and compatibility

No storage, receipt or workspace-name contract change.

Files and ownership

• ui/src/features/agent/agentActions.ts — classifiers
• ui/tests/agentActions.test.mjs — regression tests

Intentionally untouched: workflow persist, DramaBox duration, image commands, MCP, launchers.

Validation

• Date (UTC): 2026-09-09

• Base SHA: d88c85d

• Head SHA: bf94304

• Validation scope: focused

• python3 scripts/verify_clean_repo.py — PASS

• python -m compileall — N/A (no Python files in this PR)

• Focused Python tests: N/A

• cd ui && npm run i18n:check — not required; no copy catalog change

• UI tests: npx tsx --tsconfig tsconfig.app.json --import ./tests/setupI18n.ts --test tests/agentActions.test.mjs — 73 pass / 0 fail

• eslint on changed files --max-warnings=0 — PASS

• cd ui && npm run build — pending CI

• git diff --check — PASS

• E2E/smoke checks: N/A (classifier unit coverage; no real generation)

Code quality

• Score: pending CI
• Complexity trend: pending CI
• Production LOC trend: pending CI
• Regression versus main: pending CI
• Ratchet: pending CI

CI and review

• CI of this HEAD: pending

• Independent agent review of this HEAD: pending

• Human merge click (operational, not code review): pending

• Reviewed at current HEAD

• Earlier review; HEAD has changed since (stale)

• Pending

• Unavailable

Coste de la tarea

• Tests simulados: 73 UI (0 tokens externos)
• Tests reales: N/A
• Llamadas LLM externas: 0
• Tokens de prompt: N/A
• Tokens de respuesta: N/A
• Tokens totales: N/A
• Generaciones de imágenes/audio/vídeo: 0
• Tiempo transcurrido: N/A
• Proveedores/modelos: N/A

Notes and limitations

• «¿Puedo cancelar la generación?» is treated as capability/how-to. «¿Puedes cancelar la generación?» remains a cancel command.
• «stop the video» still hits the example-turn helper before cancel (pre-existing; not changed here).
• Residual from the feat(studio): share durable image commands with Wizard and MCP #275 audit, not in this PR: one corrupt localStorage pending image command can hide every other pending recovery row.

Follow-up work

• Isolate corrupt pending image-command rows in pendingImageGenerationCommands so one bad key cannot hide the rest.

Checklist

• The executive summary is understandable without reading the code.
• The detailed Summary/Overview has not been removed or shortened.
• Tests and their actual results are recorded.
• Generated assets, secrets and local-only files are not committed.
• Required CI and Cursor/Bugbot review are complete, or the PR is clearly marked as waiting for them.

  

[github-actions]

PR Review — Loreframe Studio

Risk: low
Scope: 2 file(s); +96/-1; React UI

Automated review from scripts/analyze_pr.py. This is a heuristic pass (no LLM) so humans still own the merge decision.

Findings

• low — UI changed — rebuild before merge
  Run cd ui && npm run build (CI already does this). Pinokio Update rebuilds for end users; keep ui/dist untracked.

Changed files

• modified: ui/src/features/agent/agentActions.ts, ui/tests/agentActions.test.mjs

CONTRIBUTING checklist

• python scripts/verify_clean_repo.py
• python -m compileall -q app/services app/launch.py scripts
• cd ui && npm run build if the UI changed
• No weights, CivitAI sidecars, or generated guides
• Stays local-first (no required accounts / telemetry)

Posted by the repo PR review workflow. Re-runs on each push to the PR.

[github-actions]

Code health

Quality score: 57.1/100

Higher is better. The score is a trend dashboard; the independent ratchet below remains the CI gate.

Component	Weight	Current	Change
Cyclomatic health	45%	55.4	+0.0
File concentration	25%	64.2	+0.0
Oversized-file debt	20%	40.1	+0.0
Modularity	10%	81.0	+0.0

Change vs PR base: +0.0 points.

Metric	Value
Production LOC	274,323
Production files	800
Test LOC	104,478
Functions measured	18,159
Functions complexity ≥ 15	861
Maximum complexity	674
Policy	code-health-policy-v1
HEAD	d33caf0b53ee0e9dcb01f40fe8229a2572f351e8
Base	d88c85d543100dad3469014ff0d4c7bd7b242222
UI measurement	complete

Markdown, JSON catalogs and tests are out of this table. Only app/ runtime + ui/src TS/JS count.

Most complex functions

Complexity	Where
674	app/wgp.py:7221 generate_video
379	ui/src/stores/useStore.ts:4086 Async method 'startGeneration'
355	app/_launch_runtime.py:23618 _run_generation
308	app/wgp.py:12363 generate_video_tab
268	ui/src/stores/useStore.ts:8638 Async method 'loadSettingsFromOutput'
262	ui/src/components/Sidebar/SceneAnimatorPanel.tsx:496 Function 'SceneAnimatorPanel'
258	app/services/director/planners/short_film.py:3433 ShortFilmPlanner._plan_story_driven
248	app/services/director_pipeline.py:13607 _run_video_generation
245	app/services/director_pipeline.py:7712 _run_pipeline
244	ui/src/features/agent/agentActions.ts:1150 Function 'parseAction'
226	app/services/director_pipeline.py:6541 update_comic_preview
225	ui/src/features/agent/agentActions.ts:2952 Async function 'executeAgentActions'

Trend vs baseline

Metric	Δ
Production LOC	+16
Test LOC	+79
Functions ≥ 15	+0
Maximum complexity	+0

Warnings

• production LOC increased by +16
• hotspot ui/src/features/agent/agentActions.ts increased by +16 lines

Ratchet passed.