paper-talk
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChinesePaper Talk: End-to-End Conference Talk Pipeline
Paper Talk:端到端会议演讲全流程
Workflow: from a completed paper to a conference-ready talk artifact —
slide outline, Beamer source, editable PPTX, speaker notes, full talk
script, polished visuals, and assurance checks.
Pipeline target: $ARGUMENTS
工作流:从已完成的论文到可直接用于会议的演讲成果——幻灯片大纲、Beamer源码、可编辑PPTX、演讲者备注、完整讲稿、优化后的视觉效果,以及合规检查。
流程目标:$ARGUMENTS
Assurance Ladder
合规等级阶梯
The skill takes an explicit argument; default is .
— assurance:polished| Level | Phases run | Use when |
|---|---|---|
| 0 → 1 → 2 → 5 → 6 | Internal practice talk; quick first-pass deck |
| 0 → 1 → 2 → 3 → 5 → 6 | Lab seminar, workshop, default conference talk |
| 0 → 1 → 2 → 3 → 4 → 5 → 6 | Top-venue oral / spotlight where slide claims are scrutinised; or anonymous submission |
Phase 4 (assurance checks) is opt-in via .
The lower levels skip claim / citation / anonymity audits, which is correct
when content has already been audited at the paper-writing stage.
— assurance: conference-ready该技能支持显式的参数;默认值为。
— assurance:polished| 等级 | 执行阶段 | 适用场景 |
|---|---|---|
| 0 → 1 → 2 → 5 → 6 | 内部演练演讲;快速生成第一版幻灯片 |
| 0 → 1 → 2 → 3 → 5 → 6 | 实验室研讨会、工作坊、常规会议演讲 |
| 0 → 1 → 2 → 3 → 4 → 5 → 6 | 顶级会议口头报告/重点展示(幻灯片论点会被严格审查);或匿名投稿演讲 |
阶段4(合规检查)需通过主动开启。
低等级会跳过论点/引用/匿名性审计,这在论文撰写阶段已完成内容审核的场景下是合理的。
— assurance: conference-readyHard Invariants
核心不变规则
These are non-negotiable across all phases:
- Original paper is read-only. This workflow consumes the paper directory; it never modifies or other paper artifacts.
paper/main.tex - Original deck is preserved. produces a baseline deck;
/paper-slideswrites a/slides-polishversioned copy. The original Beamer + PPTX are never overwritten._polished - Speaker notes are byte-stable. Polish must not change content. Phase 4 verifies this.
slide.notes_slide - No new content anywhere in the pipeline. All slide text, speaker notes, talk script, Q&A answers, claims, numbers, citations, URLs, author names, affiliations, anonymity placeholders, and experiment results must be either paper-grounded (extracted from artefacts) or explicitly user-provided. The pipeline never invents content during outline, build, polish, audit, or export. Phase-4 anonymity scan + claim audit verify this end-to-end.
PAPER_DIR/ - No slide reordering. Add / drop / reorder requires explicit user flags.
- Cross-model independence. Per-page Codex calls in use fresh threads (no
/slides-polish). Seecodex-reply.../shared-references/reviewer-independence.md - Anonymity fail-closed. If any audit (or any Codex fix proposal) would replace a placeholder with a real title / count / URL, the workflow halts and surfaces the proposal for human review. See .
../shared-references/experiment-integrity.md - Style references are guidance, not text source. A PDF or
— reference:preset informs visual weight and structural rhythm; never copy prose, examples, slide titles, or speaker-note text from the reference.— style: - Final report cannot be unless required audits pass. Phase 6 verifies and downgrades verdict if audits fail.
conference-ready - is invariant across all
reasoning_effort: xhighlevels for any Codex call invoked by sub-skills.effort
以下规则在所有阶段均不可违反:
- 原始论文为只读状态。该工作流仅读取论文目录;绝不会修改或其他论文产物。
paper/main.tex - 原始幻灯片保留。生成基线幻灯片;
/paper-slides会写入带/slides-polish后缀的版本副本。原始Beamer + PPTX文件绝不会被覆盖。_polished - 演讲者备注字节级稳定。优化阶段不得修改内容。阶段4会对此进行验证。
slide.notes_slide - 流程中不得生成新内容。所有幻灯片文本、演讲者备注、讲稿、问答内容、论点、数据、引用、URL、作者姓名、机构信息、匿名占位符及实验结果,必须要么来自论文(从产物中提取),要么由用户明确提供。在大纲生成、幻灯片构建、优化、审计或导出阶段,流程绝不会自行生成内容。阶段4的匿名性扫描+论点审计会端到端验证这一点。
PAPER_DIR/ - 不得重新排序幻灯片。添加/删除/重新排序需用户明确设置参数。
- 跨模型独立。中的逐页Codex调用使用全新线程(不依赖
/slides-polish)。详见codex-reply。../shared-references/reviewer-independence.md - 匿名性检查失败即终止。若任何审计(或Codex修复建议)会将占位符替换为真实标题/数据/URL,工作流将暂停并将建议提交给人工审核。详见。
../shared-references/experiment-integrity.md - 风格参考仅作指导,而非文本来源。指定的PDF或
— reference:预设仅用于指导视觉权重和结构节奏;绝不能从参考中复制文案、示例、幻灯片标题或演讲者备注文本。— style: - 最终报告仅在通过要求的审计后才能标记为。阶段6会进行验证,若审计失败则降级结论。
conference-ready - ****在所有
reasoning_effort: xhigh等级下,对所有子技能调用的Codex请求均保持不变。effort
Constants
固定参数
- PAPER_DIR = — Source paper directory. Override via positional argument.
paper/ - OUTPUT_DIR = — Where the deck artefacts live.
slides/ - STATE_DIR = — Workflow state, audit logs, final report.
.aris/paper-talk/ - TALK_TYPE = — Default talk format. Inherited by
spotlight./paper-slides - TALK_MINUTES = 15 — Default duration. Inherited by .
/paper-slides - VENUE = — Default venue (used by
NeurIPScolor schemes when/paper-slidesis not passed).— style - ASPECT_RATIO = — Inherited by
16:9./paper-slides - STYLE_PRESET — if not passed;
genericand venue presets supported bywhy-rf./slides-polish - REFERENCE_VISUAL — Required when . The Beamer compile of this talk is an acceptable self-reference; an external academic talk PDF is preferred when the user wants visual alignment beyond defaults.
assurance ≥ polished - AUTO_PROCEED = false — Each major phase pauses for user approval. Set only when explicitly requested.
true
- PAPER_DIR = — 源论文目录。可通过位置参数覆盖。
paper/ - OUTPUT_DIR = — 幻灯片产物存储目录。
slides/ - STATE_DIR = — 工作流状态、审计日志、最终报告存储目录。
.aris/paper-talk/ - TALK_TYPE = — 默认演讲格式。由
spotlight继承。/paper-slides - TALK_MINUTES = 15 — 默认演讲时长。由继承。
/paper-slides - VENUE = — 默认会议场地(当未传递
NeurIPS参数时,— style会使用该场地的配色方案)。/paper-slides - ASPECT_RATIO = — 由
16:9继承。/paper-slides - STYLE_PRESET — 未传递时为;
generic支持/slides-polish及场地预设。why-rf - REFERENCE_VISUAL — 当时为必填项。该演讲的Beamer编译版可作为自参考;若用户希望视觉效果超出默认值,优先使用外部学术演讲PDF。
assurance ≥ polished - AUTO_PROCEED = false — 每个主要阶段都会暂停等待用户确认。仅当用户明确要求时才设置为。
true
Inputs
输入项
Discovered from and the project directory:
$ARGUMENTS- — compiled paper (
PAPER_DIR/+main.tex). The paper must already be in good shape; this workflow does not write the paper.main.pdf - — section sources for content extraction.
PAPER_DIR/sections/*.tex - — figures available for slide reuse.
PAPER_DIR/figures/ - (optional) — pre-existing outline (resume mode); skips Phase 1 generation.
slides/SLIDE_OUTLINE.md - (optional) — visual anchor for
— reference: <pdf>. If absent and the assurance level requires polish, the workflow uses the Beamer self-compile as reference./slides-polish - (optional) ,
— talk_type,— minutes,— style— see Constants.— effort
从和项目目录中获取:
$ARGUMENTS- — 已编译的论文(
PAPER_DIR/+main.tex)。论文需已完成定稿;该工作流不负责撰写论文。main.pdf - — 用于内容提取的章节源码。
PAPER_DIR/sections/*.tex - — 可复用至幻灯片的图片。
PAPER_DIR/figures/ - (可选)— 已存在的大纲(恢复模式);跳过阶段1的生成步骤。
slides/SLIDE_OUTLINE.md - (可选)—
— reference: <pdf>的视觉锚点。若未提供且合规等级要求优化,工作流会使用Beamer自编译版作为参考。/slides-polish - (可选),
— talk_type,— minutes,— style— 详见固定参数。— effort
Output Layout
输出结构
File names match 's emit contract — this workflow is a
non-renaming consumer.
/paper-slidesslides/
├── SLIDE_OUTLINE.md # Phase-1 outline (claim-first per slide)
├── main.tex # Beamer source (from /paper-slides)
├── main.pdf # compiled Beamer
├── presentation.pptx # editable PPTX (from /paper-slides; emit name fixed by /paper-slides)
├── presentation_pre_polish.pptx # snapshot before /slides-polish
├── presentation_polished.pptx # /slides-polish output (when assurance ≥ polished)
├── presentation_polished.pdf # rendered (LibreOffice or user-export)
├── speaker_notes.md # per-slide notes (from /paper-slides)
├── TALK_SCRIPT.md # full word-for-word talk script + Q&A (from /paper-slides)
└── assets/ # per-slide PNG previews
.aris/paper-talk/
├── PIPELINE_STATE.json # phase pointer, status, timestamps
├── FINAL_REPORT.md # human-readable summary at end
├── audit-input/ # Phase-4 staging copies of slide text + notes + script (so /paper-claim-audit and /citation-audit can run on slide content as a synthetic "paper")
│ ├── main.tex # \input{sections/slide_text.tex} \input{sections/notes.tex} \input{sections/script.tex}
│ ├── sections/
│ │ ├── slide_text.tex # all visible slide bullets/titles/callouts as \section{Slide K}
│ │ ├── notes.tex # speaker_notes.md → LaTeX, one \section per slide
│ │ └── script.tex # TALK_SCRIPT.md → LaTeX, one \section per slide
│ ├── references.bib → ../../../paper/references.bib # symlink to source paper bib for /citation-audit
│ ├── results/ → ../../../paper/results/ # symlink — claim audit needs real evidence
│ └── figures/ → ../../../paper/figures/ # symlink — citation audit may reference figure captions
└── audits/
├── slide_claim_audit.json # 6-state verdict per claim (PASS/WARN/FAIL/NOT_APPLICABLE/BLOCKED/ERROR)
├── citation_audit.json # citation existence / metadata
├── anonymity_scan.json # placeholder discipline check
└── export_integrity.json # page-count / aspect / blank-slide / notes-preservedThe audit JSON files follow the shared 6-state schema; see
.
../shared-references/assurance-contract.md文件名遵循的输出约定 — 该工作流为非重命名消费者。
/paper-slidesslides/
├── SLIDE_OUTLINE.md # 阶段1生成的大纲(每页以论点为核心)
├── main.tex # Beamer源码(来自/paper-slides)
├── main.pdf # 编译后的Beamer文件
├── presentation.pptx # 可编辑PPTX(来自/paper-slides;文件名由/paper-slides固定)
├── presentation_pre_polish.pptx # /slides-polish优化前的快照
├── presentation_polished.pptx # /slides-polish输出文件(当assurance ≥ polished时生成)
├── presentation_polished.pdf # 渲染后的文件(LibreOffice或用户导出)
├── speaker_notes.md # 逐页演讲者备注(来自/paper-slides)
├── TALK_SCRIPT.md # 完整逐字讲稿+问答内容(来自/paper-slides)
└── assets/ # 逐页PNG预览图
.aris/paper-talk/
├── PIPELINE_STATE.json # 阶段指针、状态、时间戳
├── FINAL_REPORT.md # 最终可读总结报告
├── audit-input/ # 阶段4的幻灯片文本+备注+讲稿暂存副本(以便/paper-claim-audit和/citation-audit将幻灯片内容作为合成"论文"进行审核)
│ ├── main.tex # \\input{sections/slide_text.tex} \\input{sections/notes.tex} \\input{sections/script.tex}
│ ├── sections/
│ │ ├── slide_text.tex # 所有可见幻灯片项目符号/标题/标注,按\\section{Slide K}组织
│ │ ├── notes.tex # speaker_notes.md转换为LaTeX格式,每页对应一个\\section
│ │ └── script.tex # TALK_SCRIPT.md转换为LaTeX格式,每页对应一个\\section
│ ├── references.bib → ../../../paper/references.bib # 指向源论文bib文件的符号链接
│ ├── results/ → ../../../paper/results/ # 符号链接 — 论点审核需要真实证据
│ └── figures/ → ../../../paper/figures/ # 符号链接 — 引用审核可能需要参考图片标题
└── audits/
├── slide_claim_audit.json # 每个论点的6种状态结论(PASS/WARN/FAIL/NOT_APPLICABLE/BLOCKED/ERROR)
├── citation_audit.json # 引用存在性/元数据检查
├── anonymity_scan.json # 占位符规范性检查
└── export_integrity.json # 页数/比例/空白幻灯片/备注保留情况检查审计JSON文件遵循通用的6状态 schema;详见。
../shared-references/assurance-contract.mdWorkflow
工作流
Phase 0: Setup
阶段0:初始化
- Validate paper directory: . If absent → halt and ask user to point to the compiled paper.
ls $PAPER_DIR/main.tex $PAPER_DIR/main.pdf - Check prerequisites:
- LaTeX:
which xelatex pdflatex latexmk - PPTX rendering: (LibreOffice headless) — required for Phase 4 export integrity check; otherwise prompt user to export PDF manually.
which soffice - PDF tools: (poppler) — required for
which pdfinfo pdftoppmPNG rendering./slides-polish - Codex MCP availability.
- python-pptx ().
python3 -c 'import pptx'
- LaTeX:
- Resolve overrides from :
$ARGUMENTS,talk_type,minutes,assurance,reference,style.effort - State init: write with
.aris/paper-talk/PIPELINE_STATE.json, timestamp, all resolved overrides.phase: 0 - Resume mode: if exists and
slides/SLIDE_OUTLINE.mdshows recent in-progress work, prompt the user — resume from last phase or start fresh.PIPELINE_STATE.json
- 验证论文目录:检查是否存在。若缺失→暂停并提示用户指定已编译的论文。
$PAPER_DIR/main.tex $PAPER_DIR/main.pdf - 检查前置依赖:
- LaTeX:
which xelatex pdflatex latexmk - PPTX渲染:(LibreOffice无头模式)—— 阶段4导出完整性检查必需;否则提示用户手动导出PDF。
which soffice - PDF工具:(poppler)——
which pdfinfo pdftoppm生成PNG预览必需。/slides-polish - Codex MCP可用性。
- python-pptx(执行验证)。
python3 -c 'import pptx'
- LaTeX:
- 解析中的覆盖参数:
$ARGUMENTS,talk_type,minutes,assurance,reference,style。effort - 初始化状态:写入,包含
.aris/paper-talk/PIPELINE_STATE.json、时间戳及所有解析后的覆盖参数。phase: 0 - 恢复模式:若存在且
slides/SLIDE_OUTLINE.md显示近期有未完成工作,提示用户——从上次阶段恢复或重新开始。PIPELINE_STATE.json
Phase 1: Slide Outline (Checkpoint)
阶段1:幻灯片大纲(检查点)
Goal: produce or accept a that the user signs off
on before any deck-building happens.
slides/SLIDE_OUTLINE.mdIf already exists, present its summary (slide
count, time budget, claim-per-slide map) to the user and ask:
slides/SLIDE_OUTLINE.md"Use existing outline (Y/n)? Modify? Regenerate?"
Otherwise, delegate to Phase-1 only (content extraction +
slide outline generation). writes the outline into its own
state; we adopt it as .
/paper-slides/paper-slidesslides/SLIDE_OUTLINE.mdThe outline must contain, per slide:
- A claim-first title (states the slide's point, not the topic).
- One main idea per slide.
- Bullet content (sentence fragments, not paragraphs).
- Figure / equation / callout note.
- Time budget (seconds).
- Transition cue.
- Speaker note seed (1-3 sentences; expanded in Phase 2).
Checkpoint: present the outline. Default behavior: pause for user
approval. Set only when the user is explicitly
running unattended.
AUTO_PROCEED = true目标:生成或接收,并在开始构建幻灯片前获得用户确认。
slides/SLIDE_OUTLINE.md若已存在,向用户展示其摘要(幻灯片数量、时间分配、每页论点映射)并询问:
slides/SLIDE_OUTLINE.md"是否使用现有大纲(Y/n)?修改?重新生成?"
否则,仅调用的阶段1(内容提取+幻灯片大纲生成)。会将大纲写入自身状态;我们将其复制为。
/paper-slides/paper-slidesslides/SLIDE_OUTLINE.md大纲需包含每页幻灯片的以下内容:
- 以论点为核心的标题(明确幻灯片的核心观点,而非仅主题)。
- 每页一个核心思想。
- 项目符号内容(短句片段,而非段落)。
- 图片/公式/标注说明。
- 时间分配(秒)。
- 过渡提示。
- 演讲者备注初稿(1-3句话;阶段2会扩展)。
检查点:展示大纲。默认行为:暂停等待用户确认。仅当用户明确要求无人值守时设置。
AUTO_PROCEED = truePhase 2: Build Baseline Deck
阶段2:构建基线幻灯片
Invoke to generate Beamer source + PPTX from the approved
outline.
/paper-slides/paper-slides "<paper-dir>" — talk_type: <T> — minutes: <N> — venue: <V> — aspect: 16:9 — notes: trueForward if the user originally passed it (writer-side; the
reviewer-side polish in Phase 3 must NOT see it — see invariant 8).
— style-ref:/paper-slides/paper-slides- +
slides/main.tex(Beamer)slides/main.pdf - (editable PPTX)
slides/presentation.pptx slides/speaker_notes.mdslides/TALK_SCRIPT.md
After return: verify the four artefacts exist. If any missing, halt with a
diagnostic. Do not rename outputs — downstream phases bind to these
exact paths.
调用根据已确认的大纲生成Beamer源码+PPTX。
/paper-slides/paper-slides "<paper-dir>" — talk_type: <T> — minutes: <N> — venue: <V> — aspect: 16:9 — notes: true若用户最初传递了参数,则转发该参数(仅用于撰写侧;阶段3的审核侧优化不得使用该参数——见核心不变规则8)。
— style-ref:/paper-slides/paper-slides- +
slides/main.tex(Beamer文件)slides/main.pdf - (可编辑PPTX)
slides/presentation.pptx slides/speaker_notes.mdslides/TALK_SCRIPT.md
返回后:验证这四个产物是否存在。若有缺失,暂停并输出诊断信息。不得重命名输出——后续阶段依赖这些精确路径。
Phase 3: Visual Polish (when assurance ≥ polished)
阶段3:视觉优化(当assurance ≥ polished时执行)
Skip when .
assurance == draftInvoke against the freshly generated PPTX, with a
reference. If the user passed , use that. Otherwise, use the
Beamer compile () as the visual reference — the Beamer is
the design intent for the same talk, so it is the correct anchor when no
external reference is available.
/slides-polish— reference:slides/main.pdf/slides-polish "slides/presentation.pptx" — reference: slides/main.pdf — style: <preset> — effort: <effort>After return: verify exists; if
rendering tools are available, also verify .
slides/presentation_polished.pptxslides/presentation_polished.pdfThe polish phase is read-only on content (see Hard Invariants). If
emits a verdict because a Codex fix proposal
would alter content, surface that block to the user and halt rather than
overriding.
/slides-polishBLOCKED当时跳过。
assurance == draft调用对刚生成的PPTX进行优化,并指定参考文件。若用户传递了,则使用该文件;否则使用Beamer编译版()作为视觉参考——Beamer是同一演讲的设计意图,因此在无外部参考时是正确的锚点。
/slides-polish— reference:slides/main.pdf/slides-polish "slides/presentation.pptx" — reference: slides/main.pdf — style: <preset> — effort: <effort>返回后:验证是否存在;若渲染工具可用,同时验证是否存在。
slides/presentation_polished.pptxslides/presentation_polished.pdf优化阶段对内容为只读(见核心不变规则)。若因Codex修复建议会修改内容而输出结论,需将该阻塞提示给用户并暂停,而非强制覆盖。
/slides-polishBLOCKEDPhase 4: Assurance Checks (when assurance == conference-ready)
阶段4:合规检查(当assurance == conference-ready时执行)
Skip when .
assurance ∈ {draft, polished}/paper-claim-audit/citation-auditpaper/main.texpaper/sections/*.texpaper/results/paper/.aris/paper-talk/audits/PASSWARNFAILNOT_APPLICABLEBLOCKEDERROR../shared-references/assurance-contract.md当时跳过。
assurance ∈ {draft, polished}/paper-claim-audit/citation-auditpaper/main.texpaper/sections/*.texpaper/results/paper/.aris/paper-talk/audits/PASSWARNFAILNOT_APPLICABLEBLOCKEDERROR../shared-references/assurance-contract.md4.0 Staging adapter
4.0 暂存适配
The staged "paper" mirrors and 's
expected layout: at the root that s ,
plus a file and + symlinks.
/paper-claim-audit/citation-auditmain.tex\input{}sections/*.tex.bibresults/figures/.aris/paper-talk/audit-input/
├── main.tex
│ # contains: \input{sections/slide_text.tex}
│ # \input{sections/notes.tex}
│ # \input{sections/script.tex}
│ # \bibliography{references}
├── sections/
│ ├── slide_text.tex # one \section{Slide K — <title>} per slide; visible bullets, titles, callout text, in-slide \cite{...}
│ ├── notes.tex # speaker_notes.md converted to LaTeX, one \section per slide
│ └── script.tex # TALK_SCRIPT.md converted to LaTeX, one \section per slide
├── references.bib → ../../../paper/references.bib # symlink to source paper bib (or copy if symlink not allowed)
├── results/ → ../../../paper/results/ # symlink — claim audit needs real evidence
└── figures/ → ../../../paper/figures/ # symlink — citation audit may reference figure captionsIf the source paper uses a different bib name (e.g., ),
mirror the actual file. If multiple bibs are used, link them all.
paper.bibThe staged "paper" exists only for the duration of Phase 4 and is removed
in Phase 6 unless the user passes .
— keep-audit-input暂存的"论文"镜像和期望的目录结构:根目录下的通过引入,加上文件及+符号链接。
/paper-claim-audit/citation-auditmain.tex\\input{}sections/*.tex.bibresults/figures/.aris/paper-talk/audit-input/
├── main.tex
│ # 内容:\\input{sections/slide_text.tex}
│ # \\input{sections/notes.tex}
│ # \\input{sections/script.tex}
│ # \\bibliography{references}
├── sections/
│ ├── slide_text.tex # 每页对应一个\\section{Slide K — <title>};包含可见项目符号、标题、标注文本及幻灯片内的\\cite{...}
│ ├── notes.tex # speaker_notes.md转换为LaTeX格式,每页对应一个\\section
│ └── script.tex # TALK_SCRIPT.md转换为LaTeX格式,每页对应一个\\section
├── references.bib → ../../../paper/references.bib # 指向源论文bib文件的符号链接(若不允许符号链接则复制)
├── results/ → ../../../paper/results/ # 符号链接 — 论点审核需要真实证据
└── figures/ → ../../../paper/figures/ # 符号链接 — 引用审核可能需要参考图片标题若源论文使用不同的bib文件名(如),则镜像实际文件。若使用多个bib文件,则全部链接。
paper.bib暂存的"论文"仅在阶段4存在,阶段6会删除,除非用户传递参数。
— keep-audit-input4.1 Slide claim audit
4.1 幻灯片论点审核
Invoke against the staged input. Scope is slide
text + speaker notes + full talk script — talks often smuggle
unsupported claims into spoken parts that the visible bullets don't show.
/paper-claim-audit/paper-claim-audit ".aris/paper-talk/audit-input"The audit emits with the shared
6-state schema. Move it to
and de-stage the path
prefix so the verdicts cite slide K rather than synthetic-paper sections.
audit-input/PAPER_CLAIM_AUDIT.json.aris/paper-talk/audits/slide_claim_audit.jsonA or verdict on any claim downgrades the Phase-6 final
verdict from to .
FAILBLOCKEDconference-readypolished调用针对暂存输入执行审核。范围覆盖幻灯片文本+演讲者备注+完整讲稿——演讲中常将未验证的论点隐藏在口头内容中,而幻灯片可见部分并未体现。
/paper-claim-audit/paper-claim-audit ".aris/paper-talk/audit-input"审核会在中输出通用6状态结论。将其移动至,并移除路径前缀,使结论指向第K页幻灯片而非合成论文章节。
audit-input/PAPER_CLAIM_AUDIT.json.aris/paper-talk/audits/slide_claim_audit.json若任何论点的结论为或,阶段6的最终结论将从降级为。
FAILBLOCKEDconference-readypolished4.2 Citation audit
4.2 引用审核
Invoke over the staged input. Verify any in
slides + notes + script via DBLP / CrossRef; flag fabricated entries.
/citation-audit\cite{...}/citation-audit ".aris/paper-talk/audit-input"Output → (6-state).
.aris/paper-talk/audits/citation_audit.json调用针对暂存输入执行审核。通过DBLP / CrossRef验证幻灯片+备注+讲稿中的所有;标记伪造的引用条目。
/citation-audit\\cite{...}/citation-audit ".aris/paper-talk/audit-input"输出→(6状态)。
.aris/paper-talk/audits/citation_audit.json4.3 Anonymity scan
4.3 匿名性扫描
When the talk is for an anonymous-submission venue or the user passed
, scan slides + notes + script for:
— anonymous- Real paper titles where placeholders should be (,
[anonymized], etc.).Withheld for anonymous review - Real author names beyond the talk's own author block.
- Real submission counts.
- Real URLs that would deanonymize.
Output → (6-state). Any
(real-content leak) blocks .
.aris/paper-talk/audits/anonymity_scan.jsonFAILconference-ready当演讲用于匿名投稿场地或用户传递参数时,扫描幻灯片+备注+讲稿以检查:
— anonymous- 应使用占位符(、
[anonymized]等)的位置出现真实论文标题。Withheld for anonymous review - 演讲作者信息之外出现真实作者姓名。
- 真实投稿数量。
- 会暴露身份的真实URL。
输出→(6状态)。任何(真实内容泄露)会阻止结论标记为。
.aris/paper-talk/audits/anonymity_scan.jsonFAILconference-readyPhase 5: Final Export + Integrity
阶段5:最终导出+完整性检查
Resolve the final PPTX path:
FINAL_PPTX = slides/presentation_polished.pptx if Phase 3 ran
= slides/presentation.pptx if assurance == draft (Phase 3 skipped)Then:
- Recompile Beamer cleanly: (or
cd slides && latexmk -pdf -xelatex main.texfor non-CJK). Confirmpdflatexpage count matches outline slide count.main.pdf - Render → PDF via
FINAL_PPTX. If unavailable, prompt user.soffice --headless --convert-to pdf <FINAL_PPTX> -outdir slides/ - Export integrity check → (6-state):
.aris/paper-talk/audits/export_integrity.json- PPTX-PDF page count == Beamer-PDF page count.
- Aspect ratio == declared (16:9 default).
- No fully-blank slide.
- Speaker notes preserved on every slide that had them in the baseline. When Phase 3 ran, sha256 match against 's notes; when Phase 3 skipped, against
presentation_pre_polish.pptx's own notes (no-op).presentation.pptx - Embedded fonts are appropriate for projector use (e.g., Calibri / Helvetica Neue / PingFang SC for Chinese).
If any integrity check fails, downgrade the Phase-6 verdict and surface
specifics to the user.
确定最终PPTX路径:
FINAL_PPTX = slides/presentation_polished.pptx 若阶段3已执行
= slides/presentation.pptx 若assurance == draft(阶段3跳过)然后执行:
- 重新编译Beamer:(非CJK文档使用
cd slides && latexmk -pdf -xelatex main.tex)。确认pdflatex页数与大纲幻灯片数量一致。main.pdf - 将FINAL_PPTX渲染为PDF:通过执行。若工具不可用,提示用户手动操作。
soffice --headless --convert-to pdf <FINAL_PPTX> -outdir slides/ - 导出完整性检查→(6状态):
.aris/paper-talk/audits/export_integrity.json- PPTX导出PDF的页数 == Beamer PDF的页数。
- 比例 == 声明的比例(默认16:9)。
- 无完全空白的幻灯片。
- 所有基线中包含备注的幻灯片,备注均已保留。若阶段3已执行,需与的备注进行sha256匹配;若阶段3跳过,则与
presentation_pre_polish.pptx自身的备注匹配(无操作)。presentation.pptx - 嵌入字体适合投影使用(如中文使用Calibri / Helvetica Neue / PingFang SC)。
若任何完整性检查失败,阶段6的结论将降级,并向用户展示具体失败细节。
Phase 6: Final Report
阶段6:最终报告
Write with:
.aris/paper-talk/FINAL_REPORT.md- Verdict: |
draft|polished(downgraded if audits failed).conference-ready - Artefact paths: Beamer source / PDF, baseline PPTX, polished PPTX, polished PDF, notes, script.
- Slide count, time budget vs target.
- Audit summary (one line per audit run + PASS / WARN / FAIL).
- Open warnings: any unresolved Codex polish notes the user should review by hand.
- Next steps: e.g., "drop in real QR images on slide N", "verify HF Daily Papers screenshot", "rehearse to confirm 25-min budget".
A final verdict requires:
conference-ready- Phases 0-5 all completed without halt.
- All Phase-4 audit JSONs returned (or
PASSwhen the audit's content detector did not match — e.g., no citations on slides). AnyNOT_APPLICABLE,WARN,FAIL, orBLOCKEDdowngrades the verdict.ERROR - Phase-5 export integrity returned .
PASS - No anonymity-scan when anonymous submission is declared.
FAIL
Otherwise the report emits the highest passing level ( or
) and itemises why was not granted, citing the
specific audit JSON and the failing verdict line.
polisheddraftconference-ready写入,包含:
.aris/paper-talk/FINAL_REPORT.md- 结论:|
draft|polished(若审计失败则降级)。conference-ready - 产物路径:Beamer源码/PDF、基线PPTX、优化后PPTX、优化后PDF、备注、讲稿。
- 幻灯片数量、时间分配与目标对比。
- 审计总结(每次审核一行 + PASS / WARN / FAIL)。
- 未解决警告:用户需手动查看的Codex优化未决建议。
- 后续步骤:例如"在第N页插入真实二维码图片"、"验证HF Daily Papers截图"、"排练以确认25分钟时长"。
最终结论需满足:
conference-ready- 阶段0-5全部完成且未暂停。
- 阶段4所有审计JSON均返回(或当审计内容检测器未匹配时返回
PASS——如幻灯片无引用)。任何NOT_APPLICABLE、WARN、FAIL或BLOCKED都会降级结论。ERROR - 阶段5导出完整性检查返回。
PASS - 当声明为匿名投稿时,匿名性扫描无结论。
FAIL
否则报告将输出最高通过等级(或),并逐项说明未达到的原因,引用具体审计JSON及失败结论行。
polisheddraftconference-readyEffort Levels
工作量等级
See .
../shared-references/effort-contract.md | Behavior |
|---|---|
| Forwarded to sub-skills. |
| Standard pipeline. |
| |
| |
reasoning_effort: xhighassuranceefforteffort: lite, assurance: conference-ready详见。
../shared-references/effort-contract.md | 行为 |
|---|---|
| 传递给子技能。 |
| 标准流程。 |
| |
| |
reasoning_effort: xhighassuranceefforteffort: lite, assurance: conference-readyParameter Pass-Through
参数传递示例
/paper-talk "paper/" — talk_type: oral — minutes: 25 — assurance: conference-ready \
— reference: ./refs/why_rf_2025.pdf — style: why-rf — effort: maxForwarded to:
- (Phase 2):
/paper-slides,paper-dir,talk_type,minutes,venue,aspect,notesif writer-side.style-ref - (Phase 3): polished pptx target,
/slides-polish,reference,style.effort - (Phase 4.1): scoped to slide deck artefacts (not paper).
/paper-claim-audit - (Phase 4.2): scoped to slide-deck-only citations.
/citation-audit - Anonymity scan (Phase 4.3): inline tool, no external skill.
/paper-talk "paper/" — talk_type: oral — minutes: 25 — assurance: conference-ready \\
— reference: ./refs/why_rf_2025.pdf — style: why-rf — effort: max传递给:
- (阶段2):
/paper-slides,paper-dir,talk_type,minutes,venue,aspect, 若为撰写侧则传递notes。style-ref - (阶段3):优化后的PPTX目标、
/slides-polish,reference,style。effort - (阶段4.1):范围限定为幻灯片产物(而非论文)。
/paper-claim-audit - (阶段4.2):范围限定为幻灯片中的引用。
/citation-audit - 匿名性扫描(阶段4.3):内置工具,无需调用外部技能。
When NOT to Use
不适用于以下场景
- The paper is not yet compiled or claims are not stable. Run first.
/paper-writing - The user wants a poster, not a talk. Use .
/paper-poster - The user already has a finished deck and only needs visual polish. Use directly —
/slides-polishwould needlessly rebuild./paper-talk - The talk content is unrelated to a paper (general lecture, demo). The orchestration assumes a paper-grounded talk; for ad-hoc decks, use directly with manual outline.
/paper-slides
- 论文尚未编译或论点未稳定。请先运行。
/paper-writing - 用户需要海报而非演讲。请使用。
/paper-poster - 用户已有完成的幻灯片,仅需视觉优化。请直接使用——
/slides-polish会不必要地重新构建幻灯片。/paper-talk - 演讲内容与论文无关(通用讲座、演示)。该流程假设演讲基于论文;对于临时幻灯片,请直接使用并手动提供大纲。
/paper-slides
Prior Skill Relationship
与其他技能的关联
- Composes ,
/paper-slides,/slides-polish,/paper-claim-audit./citation-audit - Does not call by default — that is upstream intellectual stress-test, not deck QA. Users who want talk-story stress-test before slide build can run
/kill-argumentfirst./kill-argument paper/ - Sister workflow to (paper) and
/paper-writing(poster)./paper-poster
- 整合了,
/paper-slides,/slides-polish,/paper-claim-audit。/citation-audit - 默认不会调用——该技能属于上游观点压力测试,而非幻灯片质量保证。若用户希望在构建幻灯片前对演讲内容进行压力测试,可先运行
/kill-argument。/kill-argument paper/ - 与(论文)和
/paper-writing(海报)为姐妹工作流。/paper-poster
Empirical Origin
经验来源
This workflow generalises a 30+ iteration polish run on a Chinese-spoken
academic conference talk (May 2026). The convergent learning was that talk
preparation has the same shape as paper preparation — plan, build, polish,
audit, export, report — and benefits from the same assurance ladder. The
per-page Codex polish pass (Phase 3) is the single most expensive but
highest-value step, and is hidden behind the default so users
get it without thinking about it.
polished该工作流基于2026年5月一场中文学术会议演讲的30+次迭代优化总结而来。核心发现是演讲准备与论文准备的流程一致——规划、构建、优化、审计、导出、报告——并可从相同的合规等级阶梯中获益。逐页Codex优化步骤(阶段3)是成本最高但价值最大的环节,因此默认启用等级,让用户无需额外设置即可获得该功能。",
polished