AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He
docs(quick-generate): define Quick mode by decision rights
hugohe32c8f02d77fbc
diff --git a/README.md b/README.mdindex 9175f874..4f41cbaa 100644--- a/README.md+++ b/README.md@@ -316,7 +316,7 @@ AI: Sure. Let's confirm the design spec:The AI handles everything — content analysis, visual design, SVG generation, and PPTX export.-> **Output:** The SVG pipeline has one PPTX converter: it reads `svg_output/` and writes a directly editable native DrawingML deck to `exports/<name>_<timestamp>.pptx`. The default Generate flow runs `finalize_svg.py` and produces self-contained previews in `svg_final/`; PowerPoint's manual **Convert to Shape** command is outside the supported contract. Explicit [quick generation](./skills/ppt-master/workflows/profiles/quick-generate.md) still converts sources, researches factual gaps, and prepares required images, icons, formulas, and resource manifests when needed. The current agent makes the content, page, visual, and resource decisions in context, skips Strategist, confirmation, `design_spec.md`, `spec_lock.md`, and `finalize_svg.py`, then hand-authors the SVG pages, passes the lockless Quick final quality check, and exports the final PPTX. Ordinary export capabilities remain available as needed, including native chart/table replacement, notes, motion, narration, and diagnostics; notes, custom object animation, and narration start off, and the agent may enable them when the request or deck needs them. A default-path Quick export writes the normal postflight report and snapshots `svg_output/` to `backup/<timestamp>/svg_output/`; an explicit output path keeps the ordinary no-backup behavior. By default charts and tables export as individually editable SVG-derived DrawingML shapes, which prioritize cross-app visual consistency. Pass `--native-charts-and-tables` to replace eligible groups with PowerPoint-native Chart/Table objects backed by data, which provide **Edit Data** and object-specific controls but may render differently across apps; this variant is saved as `exports/<name>_<timestamp>_native_charts_tables.pptx`. Both chart/table export variants are editable—the distinction is the PowerPoint object model, not editability itself.+> **Output:** The SVG pipeline has one PPTX converter: it reads `svg_output/` and writes a directly editable native DrawingML deck to `exports/<name>_<timestamp>.pptx`. The default Generate flow runs `finalize_svg.py` and produces self-contained previews in `svg_final/`; PowerPoint's manual **Convert to Shape** command is outside the supported contract. Explicit [quick generation](./skills/ppt-master/workflows/profiles/quick-generate.md) skips the Strategist analysis and the confirmation stage: whatever you state explicitly is followed, and whatever you leave unspecified the agent decides directly instead of asking. It still converts sources, researches factual gaps, and prepares required images, icons, formulas, and resource manifests when needed, skips `design_spec.md`, `spec_lock.md`, and `finalize_svg.py`, then hand-authors the SVG pages, passes the lockless Quick final quality check, and exports the final PPTX. Ordinary export capabilities remain available as needed, including native chart/table replacement, notes, motion, narration, and diagnostics; notes, custom object animation, and narration start off, and the agent may enable them when the request or deck needs them. A default-path Quick export writes the normal postflight report and snapshots `svg_output/` to `backup/<timestamp>/svg_output/`; an explicit output path keeps the ordinary no-backup behavior. By default charts and tables export as individually editable SVG-derived DrawingML shapes, which prioritize cross-app visual consistency. Pass `--native-charts-and-tables` to replace eligible groups with PowerPoint-native Chart/Table objects backed by data, which provide **Edit Data** and object-specific controls but may render differently across apps; this variant is saved as `exports/<name>_<timestamp>_native_charts_tables.pptx`. Both chart/table export variants are editable—the distinction is the PowerPoint object model, not editability itself.> **Already have a `.pptx` you want to reuse?** Hand the AI that deck plus your material and ask it to "fill this deck with the new content" — it fills text, table, and chart data into your existing design and exports only the pages you pick, staying natively editable. See the [FAQ](./docs/faq.md) and [template-fill workflow](./skills/ppt-master/workflows/template-fill-pptx.md).diff --git a/README_CN.md b/README_CN.mdindex 795fe7a7..22851962 100644--- a/README_CN.md+++ b/README_CN.md@@ -325,7 +325,7 @@ AI:好的,先确认设计规范:AI 全程处理——内容分析、视觉设计、SVG 生成、PPTX 导出。-> **输出说明:** SVG 管线统一由项目转换器读取 `svg_output/`,生成可直接编辑的原生 DrawingML `.pptx`,保存至 `exports/<name>_<timestamp>.pptx`。默认 Generate 流程会运行 `finalize_svg.py` 并生成自包含预览 `svg_final/`;PowerPoint 手工“转换为形状”不在支持范围。用户可显式启用[快速生成模式](./skills/ppt-master/workflows/profiles/quick-generate.md):仍按需转换来源、研究事实缺口,并准备图片、图标、公式和资源 manifest;当前 Agent 在上下文中自动决定内容、页结构、视觉和资源,跳过 Strategist、确认、`design_spec.md`、`spec_lock.md` 与 `finalize_svg.py`,随后按规范手写 SVG,通过无锁的 Quick 最终质量检查,再导出最终 PPTX。原生图表 / 表格替换、讲稿、动效、旁白和诊断等普通导出能力仍可按需使用;讲稿、自定义对象动画和旁白默认关闭,Agent 可在用户要求或 deck 确有需要时启用。Quick 使用默认输出路径时仍会生成普通 postflight 报告,并把 `svg_output/` 镜像到 `backup/<timestamp>/svg_output/`;显式指定输出路径时沿用普通流程不创建备份的行为。图表和表格默认导出为 SVG 派生、可逐形状编辑的 DrawingML 对象,优先保证 PowerPoint / Keynote / WPS 间的视觉一致性;可加 `--native-charts-and-tables`,把符合合同的组替换为带数据源和对象专属编辑能力的 PowerPoint 原生 Chart/Table 对象,跨软件渲染可能略有差异,保存为 `exports/<name>_<timestamp>_native_charts_tables.pptx`。这两种图表/表格导出变体都可编辑,区别在于 PowerPoint 对象模型,而不是“能否编辑”。+> **输出说明:** SVG 管线统一由项目转换器读取 `svg_output/`,生成可直接编辑的原生 DrawingML `.pptx`,保存至 `exports/<name>_<timestamp>.pptx`。默认 Generate 流程会运行 `finalize_svg.py` 并生成自包含预览 `svg_final/`;PowerPoint 手工“转换为形状”不在支持范围。用户可显式启用[快速生成](./skills/ppt-master/workflows/profiles/quick-generate.md):它跳过策略师分析与确认阶段——你明确提出的要求照做,你没提的由 Agent 直接决定,不再回来问你;仍按需转换来源、研究事实缺口,并准备图片、图标、公式和资源 manifest,跳过 `design_spec.md`、`spec_lock.md` 与 `finalize_svg.py`,随后按规范手写 SVG,通过无锁的 Quick 最终质量检查,再导出最终 PPTX。原生图表 / 表格替换、讲稿、动效、旁白和诊断等普通导出能力仍可按需使用;讲稿、自定义对象动画和旁白默认关闭,Agent 可在用户要求或 deck 确有需要时启用。Quick 使用默认输出路径时仍会生成普通 postflight 报告,并把 `svg_output/` 镜像到 `backup/<timestamp>/svg_output/`;显式指定输出路径时沿用普通流程不创建备份的行为。图表和表格默认导出为 SVG 派生、可逐形状编辑的 DrawingML 对象,优先保证 PowerPoint / Keynote / WPS 间的视觉一致性;可加 `--native-charts-and-tables`,把符合合同的组替换为带数据源和对象专属编辑能力的 PowerPoint 原生 Chart/Table 对象,跨软件渲染可能略有差异,保存为 `exports/<name>_<timestamp>_native_charts_tables.pptx`。这两种图表/表格导出变体都可编辑,区别在于 PowerPoint 对象模型,而不是“能否编辑”。> **已有一份想复用的 `.pptx`?** 把那份 deck 连同素材给 AI,说「套模板」即可——它会把新内容(文字、表格、图表数据)填进你现有的设计,只导出你挑选的页面,且保持原生可编辑。详见 [常见问题](./docs/zh/faq.md) 与 [套模板工作流](./skills/ppt-master/workflows/template-fill-pptx.md)。diff --git a/docs/faq.md b/docs/faq.mdindex e046082f..06cee78f 100644--- a/docs/faq.md+++ b/docs/faq.md@@ -201,29 +201,45 @@ A typical 10–15 page presentation takes about **10–20 minutes** with a fastIf generation feels slow, check your model's token throughput. The bottleneck is usually the model's output speed, not the scripts.-## Q: Can I generate directly without the Strategist phase?--Yes. Explicitly request **quick generation**. The Generate route then uses the-[`quick-generate` profile](../skills/ppt-master/workflows/profiles/quick-generate.md):-source conversion and research for identified factual gaps still run when-needed, but the current agent decides the content, page structure, visual-system, and resource roster in active context without invoking Strategist,-confirmation, `design_spec.md`, or `spec_lock.md`. It also skips-`finalize_svg.py`, so Quick creates no `svg_final/` preview.--The profile still prepares every resource the deck needs: supplied or extracted-images, AI/web/sliced images, project icons, rendered formulas, and the required-manifests or provenance records. After preparation, the current agent-hand-authors `svg_output/` to the shared standards, runs the lockless Quick final-quality checker, fixes every blocking error, and only then exports the final-PPTX. Ordinary exporter capabilities remain available as needed, including-native chart/table replacement, notes, motion, narration, and diagnostics.-Notes, custom object animation, and narration start off; the agent may enable-them when the request or deck needs them, without opening a confirmation flow.-A default-path export writes the normal postflight report and snapshots+If what you want is less process rather than a different model, explicitly ask+for quick generation: it skips the Strategist analysis and the confirmation+stop, so the planning phase costs nothing, but per-page SVG authoring takes the+same time. See the next question, "I don't want to confirm a design spec first+— can I generate directly?".++## Q: I don't want to confirm a design spec first — can I generate directly?++Yes. Explicitly request **quick generation**, and the Generate route uses the+[`quick-generate` profile](../skills/ppt-master/workflows/profiles/quick-generate.md).++**What it skips is the Strategist analysis, the `design_spec.md` /+`spec_lock.md` artifacts, and the staged confirmation stop: whatever you state+explicitly is followed, and whatever you leave unspecified the current agent+decides directly and continues, without coming back for approval.** State+nothing, and the agent decides everything. It also skips `finalize_svg.py`, so+Quick creates no `svg_final/` preview.++It does not skip preparation. Source conversion, research on identified factual+gaps, and every resource the deck needs still run when required: supplied or+extracted images, AI/web/sliced images, project icons, rendered formulas, and+the required manifests or provenance records. If a required asset is not ready,+it still stops and asks you for it instead of substituting unrelated material.+After preparation, the current agent hand-authors `svg_output/` to the shared+standards, runs the lockless Quick final quality checker, fixes every blocking+error, and only then exports the final PPTX.++Ordinary exporter capabilities remain available as needed, including native+chart/table replacement, notes, motion, narration, and diagnostics. Notes,+custom object animation, and narration start off; the agent may enable them when+the request or deck needs them, without opening a confirmation flow. A+default-path export writes the normal postflight report and snapshots`svg_output/` under `backup/`; an explicit output path keeps the ordinaryno-backup behavior. Page count alone neither activates nor blocks quick-generation. This is a planning shortcut, not a wall-clock or+generation.++Because the whole planning phase no longer happens, token usage is materially+lower than the default flow; per-page SVG authoring is the dominant cost of a+run and it does not shrink. So this is a planning shortcut, not a wall-clock ordefault-quality-equivalence promise.## Q: Will long decks blow out the context window in one shot?diff --git a/docs/getting-started.md b/docs/getting-started.mdindex 49cc5798..2b4d747e 100644--- a/docs/getting-started.md+++ b/docs/getting-started.md@@ -68,7 +68,26 @@ The whole loop is three steps. Install first — you only need Python; see [Quic```3. **Get an editable `.pptx`** at `exports/<name>_<timestamp>.pptx` — real DrawingML shapes, text boxes, and charts you can click and edit in PowerPoint, Keynote, WPS, or LibreOffice.-Before it starts, the AI confirms a short design spec (template, format, page count, …); from there it handles content analysis, layout, image acquisition, SVG generation, and export — the core loop everything else builds on.+Before it starts, the AI confirms a short design spec (template, format, page count, …); from there it handles content analysis, layout, image acquisition, SVG generation, and export — the core loop everything else builds on. To skip the confirmation step, see [Quick mode](#quick-mode) below.++---++## Quick mode++The default flow confirms a short design spec with you first. To skip that step, explicitly ask for **quick generation**:++```+You: Quickly generate a deck from sources/report.pdf — no need to confirm with me+You: Turn this into a deck, skip the confirmation, about 8 pages, dark corporate look+```++**Whatever you state explicitly is followed; whatever you leave unspecified the agent decides directly, without coming back to ask.** The page count and the look in the second example still hold — quick mode drops the round trip, not your say. State nothing, and the agent decides everything.++It does not skip preparation: source conversion, research on identified factual gaps, and image / icon / formula preparation still run as needed. If a required asset is missing, it still stops and asks you for it instead of substituting unrelated material.++Because the whole planning phase no longer happens (Strategist analysis, `design_spec.md` / `spec_lock.md` authoring, the staged confirmation round trip), token usage is materially lower than the default flow. Per-page SVG authoring is unchanged, so this is neither a wall-clock promise nor a promise of quality equivalence with the default flow.++Full guide → [quick-generate profile](../skills/ppt-master/workflows/profiles/quick-generate.md)---diff --git a/docs/roadmap.md b/docs/roadmap.mdindex 561a979c..9c803670 100644--- a/docs/roadmap.md+++ b/docs/roadmap.md@@ -85,10 +85,15 @@ The narrow exception is the `data-pptx-replace-with` marker: independently plannIn the cost / speed / quality triangle this project picks **quality**. ~20 minutes for a high-quality PPTX is the current reasonable point.Will do: indirect improvements via prompt slimming / cache hit rate.-Explicit `quick-generate` is a user-selected workflow shortcut, not a default-quality-preserving speed optimization: it skips Strategist, confirmation, and-the first-page gate, then authors SVG, runs one lockless final quality gate, and-exports the final PPTX. It makes no wall-clock or quality-equivalence promise.++Explicit `quick-generate` is a user-selected workflow shortcut: it skips+Strategist, confirmation, and the first-page gate, then authors SVG, runs one+lockless final quality gate, and exports the final PPTX. Because the whole+planning phase no longer happens — the Strategist reference load, the+`design_spec.md` / `spec_lock.md` artifacts, and the staged confirmation round+trip — its token cost disappears with it, while per-page SVG authoring is+unchanged. It is not a default quality-preserving speed optimization, and it+makes no wall-clock or quality-equivalence promise.The default Generate pipeline continues to prefer quality over speed.diff --git a/docs/technical-design.md b/docs/technical-design.mdindex 90b3acdc..0feae64f 100644--- a/docs/technical-design.md+++ b/docs/technical-design.md@@ -160,7 +160,7 @@ Use this table before reasoning about implementation details. Most failed runs s|---|---|---|| Topic only, or supplied material lacks facts required by the requested outcome | Generate PPTX + `topic-research` inside Step 1 | topic-only research starts immediately; source-backed research follows conversion/read and fills only identified factual gaps || Source files or conversation text, deck structure may be rethought | Generate PPTX | Strategist may split, merge, drop, reorder, and redesign |-| Explicit quick generation | Generate PPTX + `quick-generate` profile | convert/read sources, research factual gaps, and prepare required resources as needed; the current agent decides content, pages, visuals, and resources in active context, skips Strategist/confirmation/spec/lock/finalize, hand-authors SVG, passes one lockless final gate, and exports the final PPTX |+| Explicit quick generation | Generate PPTX + `quick-generate` profile | convert/read sources, research factual gaps, and prepare required resources as needed; explicit user requirements are followed and the current agent decides every remaining content, page, visual, and resource question in active context, skips Strategist/confirmation/spec/lock/finalize, hand-authors SVG, passes one lockless final gate, and exports the final PPTX || PPTX as source material, user allows a new story/page structure | Generate PPTX via `ppt_to_md` + `pptx_intake` | PPTX identity/geometry are facts and candidates, not replica constraints || Raw PPTX template plus new material/topic | Fill Native PPTX (`template-fill-pptx`) | clone/fill native slides; no SVG generation || Existing PPTX, preserve page count/order/wording 1:1, improve layout | Generate PPTX + `beautify-pptx` profile | regenerate through SVG; content and pagination are locked |@@ -191,10 +191,11 @@ Post-processing scripts convert supported SVG vector elements to DrawingML. Text`quick-generate` retains the source-understanding and resource-preparation workneeded by the deck, but skips the separate Strategist planning/confirmation-phase, first-page gate, and `finalize_svg.py`. The current agent makes those-content, page, visual, and resource decisions automatically in active context,-then authors under the shared SVG standards, runs one lockless final quality-gate, and uses the same DrawingML converter and postflight.+phase, first-page gate, and `finalize_svg.py`. The current agent follows every+explicit user requirement and makes the remaining content, page, visual, and+resource decisions automatically in active context, then authors under the+shared SVG standards, runs one lockless final quality gate, and uses the same+DrawingML converter and postflight.---@@ -339,7 +340,7 @@ These invariants are stronger than ordinary implementation preferences. If a cha|---|---|| `sources/` content-type files are the Generate content contract | text, tables, and chart values come from content-type files in `sources/` (Markdown is primary, but `.txt` / `.csv` / `.json` / `.yaml` / … count too); known sidecars (`*.conversion_profile.json`, `*_files/image_manifest.json`) are excluded || `analysis/` stores machine facts, not design contracts | `source_profile.json` and intake artifacts inform Strategist in the default pipeline and the current agent in `quick-generate`; they do not lock page count/order except in workflows that say so |-| `design_spec.md` explains the design; `spec_lock.md` executes it in the default pipeline | both remain owning artifacts there; `quick-generate` persists neither, and the current agent keeps its content, page, visual, and resource decisions in active context |+| `design_spec.md` explains the design; `spec_lock.md` executes it in the default pipeline | both remain owning artifacts there; `quick-generate` persists neither, and the current agent follows explicit user requirements while keeping every remaining content, page, visual, and resource decision in active context || Planning context is retained until invalidated | continuous execution reuses the complete Design Spec, lock, and triggered references; fresh/resumed/restarted or compacted execution reloads them once || `page-context` is on demand | the read-only projector supports diagnostics, deterministic routing checks, and optional usage telemetry; it is not a pre-page gate || `svg_output/` is the only hand-authored SVG directory | quality checks, manual edits, re-export, and `update_spec.py` target authored source |diff --git a/docs/zh/faq.md b/docs/zh/faq.mdindex 79a72451..c5fe3dd8 100644--- a/docs/zh/faq.md+++ b/docs/zh/faq.md@@ -202,11 +202,19 @@ python3 skills/ppt-master/scripts/svg_to_pptx.py <project> -a auto --animation-t如果感觉生成很慢,检查一下模型的 token 吞吐速度。瓶颈通常在模型的输出速度,而不是脚本本身。-## Q: 可以跳过策略师阶段直接生成吗?+如果你想要的是少走流程、而不是换模型,可以显式要求快速生成:它跳过策略师分析与确认停顿,规划阶段的开销随之消失,但逐页 SVG 生成的时间不变。见下一问「不想先确认设计规格,能直接生成吗?」。-可以。请显式要求**快速生成**。Generate 路线会启用 [`quick-generate` profile](../../skills/ppt-master/workflows/profiles/quick-generate.md):仍按需转换来源并研究已识别的事实缺口,但当前 Agent 会在上下文中自动决定内容、页结构、视觉系统和资源清单,不进入 Strategist、确认、`design_spec.md` 或 `spec_lock.md`;同时跳过 `finalize_svg.py`,因此不生成 `svg_final/` 预览。+## Q: 不想先确认设计规格,能直接生成吗?-该 profile 仍会按需备齐生成 deck 所需的资源:用户提供或源文件抽取的图片、AI / 网络 / 切片图片、项目图标、渲染公式,以及对应的必要 manifest 或来源记录。备料完成后,当前 Agent 按共享规范手写 `svg_output/`,运行无锁的 Quick 最终质量检查并修复所有阻塞错误,之后才导出最终 PPTX。原生图表 / 表格替换、讲稿、动效、旁白和诊断等普通导出能力仍可按需使用;讲稿、自定义对象动画和旁白默认关闭,Agent 可在用户要求或 deck 确有需要时自动启用,不会打开确认流程。使用默认输出路径时会生成普通 postflight 报告,并把 `svg_output/` 备份到 `backup/`;显式指定输出路径时沿用普通流程不创建备份的行为。页数本身既不会自动触发,也不会阻止快速生成。这是规划阶段的短路,不承诺具体耗时,也不承诺与默认流程质量等价。+可以。请显式要求**快速生成**,Generate 路线会启用 [`quick-generate` profile](../../skills/ppt-master/workflows/profiles/quick-generate.md)。++**它跳过的是策略师分析、`design_spec.md` / `spec_lock.md` 落盘和分步确认停顿:你明确提出的要求照做;你没提的,当前 Agent 直接决定并继续,不再回来征求同意。** 什么都不提,才是全部由 Agent 决定。它同时跳过 `finalize_svg.py`,因此不生成 `svg_final/` 预览。++它不跳过备料:来源转换、已识别事实缺口的研究,以及生成 deck 所需的资源仍按需准备——用户提供或源文件抽取的图片、AI / 网络 / 切片图片、项目图标、渲染公式,以及对应的必要 manifest 或来源记录。必需素材未就绪时它仍会停下来跟你要,不会拿无关材料顶替。备料完成后,当前 Agent 按共享规范手写 `svg_output/`,运行无锁的 Quick 最终质量检查并修复所有阻塞错误,之后才导出最终 PPTX。++原生图表 / 表格替换、讲稿、动效、旁白和诊断等普通导出能力仍可按需使用;讲稿、自定义对象动画和旁白默认关闭,Agent 可在用户要求或 deck 确有需要时自动启用,不会打开确认流程。使用默认输出路径时会生成普通 postflight 报告,并把 `svg_output/` 备份到 `backup/`;显式指定输出路径时沿用普通流程不创建备份的行为。页数本身既不会自动触发,也不会阻止快速生成。++由于整个规划阶段不再发生,token 消耗明显低于默认流程;逐页 SVG 生成是一次 run 的主要开销,这部分并不减少。因此它是规划阶段的短路,不承诺具体耗时,也不承诺与默认流程质量等价。## Q: 长 PPT 一次生成会不会上下文爆掉?diff --git a/docs/zh/getting-started.md b/docs/zh/getting-started.mdindex de589388..32174964 100644--- a/docs/zh/getting-started.md+++ b/docs/zh/getting-started.md@@ -68,7 +68,26 @@ Create Template 会分析参考材料,确认结果属于 Brand、Layout 还是```3. **拿回可编辑的 `.pptx`**,位于 `exports/<名称>_<时间戳>.pptx` —— 真正的 DrawingML 形状、文本框、图表,在 PowerPoint / Keynote / WPS / LibreOffice 里点开就能改。-开始前 AI 会先确认一份简短的设计规格(模板、格式、页数……);之后内容分析、排版、配图、SVG 生成、导出都由它完成——这就是其它能力围绕的核心环节。+开始前 AI 会先确认一份简短的设计规格(模板、格式、页数……);之后内容分析、排版、配图、SVG 生成、导出都由它完成——这就是其它能力围绕的核心环节。不想走确认这一步,见下方[快速模式](#快速模式)。++---++## 快速模式++默认流程会先跟你确认一份设计规格。不想经过这一步,就显式要求**快速生成**:++```+你:用 sources/report.pdf 快速生成一份 PPT,不用跟我确认+你:这份内容直接做成 PPT,跳过确认,8 页左右,深色商务风+```++**你明确提的照做,你没提的 AI 直接定,不再回来问你。** 第二个例子里的页数和风格照样生效——快速模式省掉的是来回确认,不是你的话语权;什么都不提,才是全部交给 AI 决定。++它不跳过备料:来源转换、事实缺口研究、图片 / 图标 / 公式准备仍按需运行。必需素材缺失时它会停下来跟你要,不会拿无关材料顶替。++整个规划阶段不再发生(策略师分析、`design_spec.md` / `spec_lock.md` 落盘、分步确认往返),因此 token 消耗明显低于默认流程;逐页 SVG 生成的开销不变,所以这既不是耗时承诺,也不承诺与默认流程质量等价。++完整说明 → [快速模式 profile](../../skills/ppt-master/workflows/profiles/quick-generate.md)---diff --git a/docs/zh/roadmap.md b/docs/zh/roadmap.mdindex 55ea65a9..a850e50c 100644--- a/docs/zh/roadmap.md+++ b/docs/zh/roadmap.md@@ -86,7 +86,7 @@ Generate PPTX 路线围绕完全可控的新形状、文字与版式创作。结会做:通过 prompt 精简 / 缓存命中率提升带来的间接改善。-显式 `quick-generate` 是用户主动选择的工作流短路,不是默认流程的保质提速:它跳过 Strategist、确认和首屏 gate,随后创作 SVG、运行一次无锁最终质量门,再导出最终 PPTX。它不承诺具体耗时,也不承诺与默认流程质量等价。+显式 `quick-generate` 是用户主动选择的工作流短路:它跳过 Strategist、确认和首屏 gate,随后创作 SVG、运行一次无锁最终质量门,再导出最终 PPTX。由于整个规划阶段不再发生——Strategist 系 reference 的加载、`design_spec.md` / `spec_lock.md` 的写入、分步确认往返——这部分 token 开销随之消失,而逐页 SVG 创作的开销不变。它不是默认流程的保质提速,也不承诺具体耗时或与默认流程质量等价。默认 Generate 流程仍坚持质量优先。diff --git a/docs/zh/technical-design.md b/docs/zh/technical-design.mdindex 9b151dfa..04118eba 100644--- a/docs/zh/technical-design.md+++ b/docs/zh/technical-design.md@@ -151,7 +151,7 @@ narration 标记。|---|---|---|| 只有主题,或现有材料缺少实现用户目标所需的事实 | Generate PPTX Step 1 内运行 `topic-research` | 只有主题时立即研究;有材料时先转换 / 阅读,只补已识别的事实缺口 || 有源文件或对话文本,deck 结构可以重想 | Generate PPTX | Strategist 可以拆分、合并、删除、重排和重设计 |-| 显式要求快速生成 | Generate PPTX + `quick-generate` profile | 按需转换 / 阅读来源、研究事实缺口并准备所需资源;当前 Agent 在上下文中决定内容、页结构、视觉与资源,跳过 Strategist / 确认 / spec / lock / finalize,手写 SVG、通过一次无锁 final gate 后导出最终 PPTX |+| 显式要求快速生成 | Generate PPTX + `quick-generate` profile | 按需转换 / 阅读来源、研究事实缺口并准备所需资源;用户明确提出的要求照做,其余内容、页结构、视觉与资源由当前 Agent 在上下文中决定,跳过 Strategist / 确认 / spec / lock / finalize,手写 SVG、通过一次无锁 final gate 后导出最终 PPTX || PPTX 作为源材料,用户允许重构故事和页结构 | Generate PPTX,经 `ppt_to_md` + `pptx_intake` | PPTX 身份和几何是事实与候选,不是复刻约束 || 原生 PPTX 模板 + 新材料 / 新主题 | Fill Native PPTX(`template-fill-pptx`) | 克隆并填充原生页面;不生成 SVG || 现有 PPTX,页数 / 页序 / 措辞 1:1 保留,只改善排版 | Generate PPTX + `beautify-pptx` profile | 通过 SVG 重新生成;内容和分页锁定 |@@ -180,7 +180,7 @@ Executor 角色逐页生成演示文稿的视觉内容,输出为 SVG 文件。**第三阶段:工程化转换**后处理脚本将受支持的 SVG 向量元素转换为 DrawingML。文本和向量形状会保持为 PowerPoint 原生对象——可点击、可编辑、可改样式;位图资源则复制为 PPT picture media,而不是把整页压平成一张图片。-`quick-generate` 保留 deck 所需的来源理解与资源准备,但跳过独立的 Strategist 规划 / 确认阶段、首屏 gate 与 `finalize_svg.py`。当前 Agent 在有效上下文中自动完成内容、页结构、视觉和资源决策,随后仍按共享 SVG 规范创作,运行一次无锁最终质量门,并使用同一个 DrawingML 转换器与 postflight。+`quick-generate` 保留 deck 所需的来源理解与资源准备,但跳过独立的 Strategist 规划 / 确认阶段、首屏 gate 与 `finalize_svg.py`。当前 Agent 照做用户明确提出的要求,并在有效上下文中自动完成其余的内容、页结构、视觉和资源决策,随后仍按共享 SVG 规范创作,运行一次无锁最终质量门,并使用同一个 DrawingML 转换器与 postflight。---@@ -323,7 +323,7 @@ CLI 支持 `--move`、`--copy` 和自动默认,但共享同一条固定的所|---|---|| `sources/` 内容型文件是 Generate 内容契约 | 文本、表格和图表数值来自 `sources/` 内容型文件(Markdown 为主,`.txt` / `.csv` / `.json` / `.yaml` 等同样计入);已知 sidecar(`*.conversion_profile.json`、`*_files/image_manifest.json`)排除在外 || `analysis/` 存机器事实,不存设计契约 | `source_profile.json` 和 intake artifact 在默认流程中辅助 Strategist,在 `quick-generate` 中辅助当前 Agent;除非工作流明确规定,否则不锁定页数 / 页序 |-| 默认流程由 `design_spec.md` 解释设计、`spec_lock.md` 执行设计 | 两者在默认流程中始终是权威产物;`quick-generate` 不落盘二者,由当前 Agent 在上下文中保留内容、页结构、视觉和资源决策 |+| 默认流程由 `design_spec.md` 解释设计、`spec_lock.md` 执行设计 | 两者在默认流程中始终是权威产物;`quick-generate` 不落盘二者,用户明确提出的要求照做,其余内容、页结构、视觉和资源决策由当前 Agent 保留在上下文中 || 规划上下文有效时持续复用 | 连续执行直接使用完整 Design Spec、lock 与已触发引用;fresh/resumed/restarted 或压缩后才重新读取一次 || `page-context` 按需调用 | 只读投影器用于诊断、确定性路由检查和可选的用量统计,不是逐页门禁 || `svg_output/` 是唯一手写 SVG 目录 | 质量检查、手工编辑、重导出和 `update_spec.py` 都面向作者源 |diff --git a/skills/ppt-master/workflows/profiles/quick-generate.md b/skills/ppt-master/workflows/profiles/quick-generate.mdindex a625e073..c3875108 100644--- a/skills/ppt-master/workflows/profiles/quick-generate.md+++ b/skills/ppt-master/workflows/profiles/quick-generate.md@@ -18,6 +18,7 @@ Page count alone never activates or blocks this profile.| Concern | Quick Generate contract ||---|---|+| Authority | Follow every explicit user requirement as stated; decide every unspecified choice directly without asking || Interaction | The current main agent decides content, design, resources, and implementation without Strategist, Confirm UI, or approval stops || Inputs | Any supported Generate input; convert/import sources and run bounded factual research when the input requires them || Resources | Prepare every project-local image, icon, formula, and required provenance/manifest artifact before the referencing SVG is authored |