WILLOSCAR

在 Markdown 中间层统一毕业论文的主线、术语、符号、指标、图表口径与章节边界，让整篇论文先在中间层收敛成“一篇论文”。

Generate a compact pool of discussion-worthy research directions from the signal table, writing `output/trace/IDEA_DIRECTION_POOL.md`.

Run a research Workflow end to end when the user requests a survey, brief, review, tutorial, or idea exploration; route an unbound goal, execute one eligible Unit at a time, and stop at checkpoints or diagnosed failures.

Internal template for creating or refactoring a skill into the repository's reference-first shape. **Trigger**: reference-first template, blueprint skill, create a reusable skill, refactor a script-heavy skill. **Use when**: you need a lean `SKILL.md`, explicit `references/`, machine-readable `assets/`, and a minimal deterministic `run.py`.

Audit one research Workspace for declared Unit outputs and Pipeline target Artifacts, writing `output/CONTRACT_REPORT.md`; use for mid-Run coverage snapshots or final delivery completeness, not deep provenance integrity.

Audit and rewrite evaluation/numeric claims to ensure they carry minimal protocol context (task + metric + constraint) and avoid underspecified model naming.

将中文毕业论文已有材料映射到“毕业论文角色”：把论文、模板、Overleaf 源稿、PDF、图表和实验材料按章节角色、研究问题和证据用途重新归位。

围绕毕业论文主线重构章节：把原论文式叙事改成学位论文式叙事，重写章节目标、内容比重、前后承接与论证方式。

规划中文毕业论文的图表、流程图和版面节奏：先决定哪些图表支撑主线，再做草图、拼版预演与图文映射。

Score an existing reader-facing deliverable and emit a deterministic PASS/FAIL scorecard plus bounded repair plan; use after synthesis, not to rewrite content.

Use when a `source-tutorial` workspace has ingested sources and needs a grounded tutorial contract before structure planning.

维护中文毕业论文的 `codex_md/question_list.md`：把本轮问题、边界、优先级、协作方案和验收口径结构化，作为整条 thesis pipeline 的控制面。

把已经在 Markdown 中间层收敛的中文毕业论文内容回写到 `chapters/*.tex` 与相关交付层文件，保持结构、图表、公式、交叉引用与章节承接一致。

Curate reader-facing survey tables for the Appendix (clean layout + high information density), using only in-scope evidence and existing citation keys.

Seed `outline/claim_evidence_matrix.md` from outline+notes; claim-matrix-rewriter rewrites it later from evidence packs.

Rewrite `outline/claim_evidence_matrix.md` as a projection/index of evidence packs (NO PROSE), so claims/axes are driven by `outline/evidence_drafts.jsonl` rather than outline placeholders.

Post-merge paper-voice gate: inspect the reader-facing draft for planner-talk and axis-label artifacts, then route fixes to the actual source.

对中文毕业论文进行编译、warning 分级、模板模式检查、数据与引用复查，并把问题回写成可继续迭代的 review checklist。

对中文毕业论文做最终润色与去 AI 味：重点处理章首导言、章末小结、贡献描述、总结与展望中的模板腔、宣传腔与 AI 口癖。

初始化中文毕业论文工作区：检查学校模板与已有材料的放置位置，明确提示当前还缺什么，建立 `codex_md/` / `claude_md/` / `tmp_layout*/` 等中间层目录，并生成材料盘点与初始工作文件。

Use when module planning and source coverage are done and the run needs writer-ready per-module packs.

Converge the screened direction pool into a small, discussion-ready shortlist, writing `output/trace/IDEA_SHORTLIST.md`.

Use when a tutorial module plan exists and the run needs an auditable module-to-source grounding file before prose.

Logic coherence pass for per-H3 section files: enforce a clear paragraph-1 thesis and surface paragraph-island risks (connector stats are diagnostic, not a quota) before merging.

Use when approved tutorial context packs exist and the run needs the final article-first tutorial deliverable.

Normalize terminology across a draft (canonical terms + synonym policy) without changing citations or meaning.

Write survey prose into per-section files under `sections/` so each unit can be QA'd independently before merging.

Generate lightweight section/subsection transitions (NO NEW FACTS) to prevent “island” subsections; outputs a transition map that merging/writing can weave in.

Run a tutorial-specific quality gate and write a PASS/FAIL report for the final tutorial deliverable.

Regression-check citation anchoring (citations stay in the same subsection) to prevent “polish drift” that breaks claim→evidence alignment.

Map paper notes + taxonomy into a signal table of tensions, missing pieces, and promising academic axes for brainstorm discussion. Writes `output/trace/IDEA_SIGNAL_TABLE.md`.

Download PDFs (when available) and extract plain text to support full-text evidence, writing `papers/fulltext_index.jsonl` and `papers/fulltext/*.txt`.

Download a small corpus of open-access arXiv survey/review PDFs about agentic systems and extract text for style learning.

Write H2 chapter lead blocks (`sections/S<sec_id>_lead.md`) that preview the chapter's comparison lens and connect its H3 subsections, without adding new facts.

Write the survey's front matter files (Abstract, Introduction, Related Work, Discussion, Conclusion) in paper voice, with high citation density and a single evidence-policy paragraph.

Multi-route literature expansion + metadata normalization for evidence-first surveys.

Polish a single H3 unit file under `sections/` into survey-grade prose (de-template + contrast/eval/limitation), without changing citation keys.

Identify survey/review papers in a retrieved set and extract taxonomy seeds into `outline/taxonomy.yml` (topics/subtopics/terminology).

Build per-H3 writer context packs (NO PROSE): merge briefs + evidence packs + anchor facts + allowed citations into a single deterministic JSONL, so drafting is less hollow and less brittle.

PRODUCES a citation-budget plan/report (`output/CITATION_BUDGET_REPORT.md`); does NOT edit the draft. Allocates an in-scope per-H3 budget (NO NEW FACTS) so drafts stop failing the global unique-citation gate; citation-injector applies it.

Audit/regression checks for the evidence-first survey pipeline: citation health, per-section coverage, placeholder leakage, and template repetition.

Bind papers to chapter-level sections first, writing `outline/section_bindings.jsonl` and `outline/section_binding_report.md`.

De-slot and harmonize paper voice across `sections/*.md` without changing meaning or citation keys.

Use when `evidence-review` has completed extraction and needs a bounded narrative synthesis.

Build a 2+ level taxonomy (`outline/taxonomy.yml`) from a core paper set and scope constraints, with short descriptions per node.

Build a final, machine-readable argument snapshot from current H3 sections and refresh section fingerprints before merge.

Compile the Beamer tutorial deck and write a build report.

APPLIES/consumes a `citation-diversifier` budget report (`output/CITATION_BUDGET_REPORT.md`) to inject *in-scope* citations into an existing draft (NO NEW FACTS), so the run passes the global unique-citation gate without citation dumps.

Generate and verify BibTeX entries from paper notes, writing `citations/ref.bib` and `citations/verified.jsonl`.

Evidence self-loop for surveys: read evidence bindings + evidence packs, then write an actionable upstream TODO plan (which stage/skill to fix) before writing more prose.

Expand and refine search keywords (synonyms, acronyms, exclusions) and update `queries.md`.

Rewrite limitation passages so the paper keeps limitations without falling into count-based slot phrases (e.g., \"Two limitations…\") across many H3s.

Use when `paper-review` needs a canonical manuscript text artifact before claim extraction.

Deterministically compact final H3 bodies to the active delivery profile's paragraph budget without deleting prose or changing citation-block order.

Remove repeated boilerplate across sections (methodology disclaimers, generic transitions, repeated summaries) while preserving citations and meaning.

Normalize cross-skill JSONL interfaces (ids + titles + citation key formats) so downstream skills do not rely on best-effort joins.

Use when an approved `evidence-review` protocol needs to be applied to a candidate pool.

Build per-subsection writing briefs (NO PROSE) so later drafting is driven by evidence and checkable comparison axes (not outline placeholders).

Use when `evidence-review` has screened includes and needs a schema-aligned extraction table.

Global consistency review for survey drafts: terminology, cross-section coherence, and scope/citation hygiene.

Rewrite subsection openers so they stop reading like a generated table-of-contents: remove \"overview/narration\" stems and reduce repeated opener cadences across H3s.

Convert a taxonomy (`outline/taxonomy.yml`) into a bullet-only outline (`outline/outline.yml`) with sections/subsections.

Planner-pass coverage + redundancy report for an outline+mapping, producing `outline/coverage_report.md` and `outline/outline_state.jsonl`.

Route an unbound research Goal to one Workflow or materialize its initial C0 kickoff block; use during initial binding, not for later checkpoint summaries or approval.

Use when `evidence-review` needs an operational protocol before screening and extraction.

Use when a `research-brief` workspace has a small paper set plus outline and needs a compact reader-facing briefing instead of a full survey.

Build or validate the source manifest for the source-tutorial pipeline from user-provided URLs/files.

Writing self-loop for surveys: run the strict section-quality gate, then rewrite only the failing `sections/*.md` files until the report is PASS.

Build a retrieval-informed chapter skeleton (`outline/chapter_skeleton.yml`) from taxonomy/core scope before stable H3 decomposition.

Project completed stage Artifacts into a concise human-review block in `DECISIONS.md`; use before a Workflow checkpoint, not for Workflow routing or approval.

Use when a review workspace has manuscript text and needs a traceable claim ledger.

Use when `paper-review` needs a claim-by-claim evidence gap report grounded in an extracted claim ledger.

Use when a tutorial module plan exists but each module still needs a verifiable teaching loop.

Review and record one pending human checkpoint in a research Workspace; use when a HUMAN Unit is blocked on an `Approve C*` decision, and never treat silence or artifact existence as approval.

Synthesize the shortlist into a discussion-ready research idea brainstorm memo, writing `output/REPORT.md`, `output/APPENDIX.md`, and `output/REPORT.json`.

Compile a LaTeX project and run basic QA (missing refs, bib errors, broken citations), producing `latex/main.pdf` and a build report.

Merge/simplify an over-fragmented outline to hit a paper-like section budget (NO PROSE): target final ToC ~6–8 H2, fewer thicker H3.

Lock an ideation run into a single-source-of-truth brainstorm brief (`output/trace/IDEA_BRIEF.md`) and a replayable multi-query plan (`queries.md`).

Use when a broad paper candidate pool needs deterministic deduplication and a stable core set.

Write structured notes for each paper in the core set into `papers/paper_notes.jsonl` (summary/method/results/limitations).

Extract per-subsection “anchor facts” (NO PROSE) from evidence packs so the writer is forced to include concrete numbers/benchmarks/limitations instead of generic summaries.

Retrieve arXiv paper metadata with keyword queries or import an offline arXiv export, and save results as JSONL (`papers/papers_raw.jsonl`).

Write `output/DRAFT.md` (or `output/SNAPSHOT.md`) from an approved outline and evidence packs, using only verified citation keys from `citations/ref.bib`.

Fill `outline/tables_index.md` from `outline/table_schema.md` + evidence packs (NO PROSE in cells; citation-backed rows).

Instantiate or update a workspace `UNITS.csv` from a selected pipeline and units template (deps/checkpoints/acceptance).

Use when `evidence-review` has an extraction table and needs lightweight risk-of-bias fields.

Use when an approved tutorial spec exists and the run needs a deterministic prerequisite graph before module planning.

Screen the direction pool with a discussion-first scoring pass, writing `output/trace/IDEA_SCREENING_TABLE.md`.

Use when `paper-review` has claims plus evidence gaps and needs the final referee-style report.

Fetch and normalize supported source-tutorial inputs into local, traceable text artifacts.

Execute exactly one eligible Unit in an existing research Workspace; use for stepwise or manual semantic execution when status, Attempt, Artifact, Manifest, checkpoint, and acceptance evidence must remain synchronized.

Build per-chapter (H2) writing briefs (NO PROSE) so the final survey reads like a paper (chapter leads + cross-H3 coherence) without inflating the ToC.

Audit-style editing pass for `output/DRAFT.md`: remove template boilerplate, improve coherence, and enforce citation anchoring.

Create per-subsection evidence packs (NO PROSE): claim candidates, concrete comparisons, evaluation protocol, limitations, plus citation-backed evidence snippets with provenance.

Write one survey-quality paragraph from evidence packs (tension → contrast → evaluation anchor → limitation).

Scaffold a LaTeX project (`latex/main.tex`, optional bibliography wiring, structure) from an existing Markdown draft.

Use when a tutorial concept DAG exists and the run needs an ordered teaching plan with objectives and outputs.

Use when `paper-review` needs overlap/delta positioning against provided related work.

Build chapter-level briefs (`outline/section_briefs.jsonl`) from chapter skeleton plus section bindings before stable H3 decomposition.

Map papers from the core set to each outline subsection and write `outline/mapping.tsv` with coverage tracking.