메뉴
Use when writing actual thesis prose for a specific section or subsection. Orchestrates research, drafting with scientific-writing principles, humanization, and personal style adaptation. Callable via /draft X.Y or /draft X.Y.Z.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Use when the user runs /acronyms, or asks to find, define, or mark up acronyms in a LaTeX thesis .tex file — e.g. converting literal acronyms like CMB, WIMP, dSph, NFW into glossaries macros and adding any missing definitions to acronyms.tex.
Core skill for the deep research and writing tool. Write scientific manuscripts in full paragraphs (never bullet points). Use two-stage process with (1) section outlines with key points using research-lookup then (2) convert to flowing prose. IMRAD structure, citations (APA/AMA/Vancouver), figures/tables, reporting guidelines (CONSORT/STROBE/PRISMA), for research papers and journal submissions.
Critical review of thesis sections or chapters. Evaluates scientific rigor, writing quality, logical flow, and citation completeness. Supports section-level (/referee X.Y) and chapter-level (/referee X) scope.
Research assistant that generates detailed, structural chapter outlines by synthesizing thesis context, local references, and NotebookLM insights.
Use when you need to resolve placeholder citations (e.g., surname_year) or missing bibliography entries in the LaTeX thesis by querying NotebookLM context and InspireHEP.
Use when saving insights from the current conversation to persistent storage, or when retrieving past insights before starting research or writing tasks.
| name | section_drafting |
| description | Use when writing actual thesis prose for a specific section or subsection. Orchestrates research, drafting with scientific-writing principles, humanization, and personal style adaptation. Callable via /draft X.Y or /draft X.Y.Z. |
Writes thesis prose for a specific section (X.Y) or subsubsection (X.Y.Z). Orchestrates NotebookLM research, drafting, and iterative refinement through a three-layer writing pipeline.
thesis references1b7df790-7858-4fc8-879c-39f41238c4aechapter_01) with chapter_outline.md and references.md already producedoutline.md (root thesis outline)source_registry skill must be called once per conversation before this skillAll prose passes through three layers in sequence:
scientific-writing (Primary Driver)writing_principles.md\cite{} naturally into prose, never as standalone listshumanizer (Anti-AI Pass)After the initial draft, check against the humanizer skill's anti-patterns:
Apply the author's voice while prioritizing clarity:
| Dimension | Style |
|---|---|
| Tone | Formal, academic, graduate-level assumed |
| Voice | Active "we" for own work; passive for universe/instruments |
| Concepts | Intuitive explanation first, then formal math |
| Transitions | Explicit sequential signposting ("Our first step…", "We now turn to…") |
| Sentences | Clear and precise. Qualify claims with conditions + citations, but prioritize clarity over complexity |
| Motivation | Physics always drives methodology — never ML-first |
| Introductions | "Funnel": broad context → specific problem → gap → "In this chapter we…" |
Style balance: The author writes naturally with longer, qualified sentences (Italian-English characteristic). Retain this personal voice but soften toward standard scientific English. Don't flatten into generic textbook prose — keep the personality — but prefer clarity when a sentence becomes too nested. Split complex sentences when it improves readability without losing voice.
Per-chapter overrides: If chapter_outline.md contains a ## Style Notes section, apply those overrides (e.g., "more pedagogical for this chapter", "assume less background"). During iteration, record any user style corrections for future sections.
Know these exist and reference if the situation calls for it:
| Skill | When it might help |
|---|---|
review | Automatically invoked in Step 9 for a critical review pass; can also be called manually via /referee X.Y |
scientific-brainstorming | Section outline feels thin or poorly motivated |
scientific-critical-thinking | Evaluating conflicting evidence in the literature |
scientific-visualization | Data plots or diagrams would strengthen the section |
scientific-schematics | Concept diagrams or flowcharts |
/draft X introduction — Chapter IntroductionDraft the untitled opening paragraphs of Chapter X (the "Section X.0" from the chapter outline). This is not a subsection intro — it covers the entire chapter at a high level.
chapter_outline.md (all sections) plus outline.md (thesis arc) to understand the chapter's scope and connections./draft X.Y, this mode does not drill into individual reference queries. It synthesises the chapter's Goal, Narrative, and Connections blocks into flowing prose.chapter_XX/sections/X.0_introduction.md./draft X.Y — Subsection LevelAsk the user which mode:
1. Intro-only mode: Draft just the introductory text of Section X.Y — the prose that appears before the first subsubsection (X.Y.1). Use when subsubsections will be drafted separately.
2. Full-section mode: Draft the entire Section X.Y including all subsubsections autonomously with self-quality gates:
Do NOT begin researching, drafting, or working on subsubsection X.Y.(Z+1) until subsubsection X.Y.Z has been fully drafted, saved to file, and passed the self-review quality gate (Step 7). No parallel drafting of multiple subsubsections. No skipping the self-review.Announce after each subsubsection: "Self-review GATE for X.Y.Z: I will not begin X.Y.(Z+1) until I have critically reviewed X.Y.Z, identified weaknesses, and revised them."
A gate violation is: starting research for the next subsubsection before the current one is saved and self-reviewed, drafting prose for multiple subsubsections in one pass, or skipping the self-criticism step.
The full flow is:
for Z in [1, 2, …, last subsubsection]:
Step 4 → Reference-guided research for X.Y.Z
Step 5 → Figure discovery for X.Y.Z
Step 6 → Draft writing (three-layer pipeline)
Step 7 → Self-review & self-criticism (MANDATORY — revise before proceeding)
Save revised draft to file
── SELF-QUALITY GATE: announce gate, confirm quality, then proceed ──
# After all subsubsections pass their self-quality gates:
Write section introduction (text before X.Y.1)
Verify inter-subsubsection transitions
Step 8 → Coherence check on the assembled section (fix issues)
Step 9 → Review pass: run the review skill, apply fixes
Step 10 → Present the revised section to the user
/draft X.Y.Z — Subsubsection LevelDraft a single subsubsection. Before writing, read existing context:
chapter_outline.md) — set up proper transitions/draft command:
X introduction → chapter introduction modeX.Y → subsection modeX.Y.Z → subsubsection modeX → directory chapter_0X/chapter_outline.mdX introduction): proceed to draft the X.0 block — read the full chapter outline and thesis connections, skip per-subsection researchX.Y): ask user for intro-only or full-section modeX.Y.Z): proceed directlyoutline.md — thesis arc, how this chapter connects to predecessors and successorschapter_outline.md — section's Goal, Narrative, key points, transitions, style notesreferences.md — key references with provenance (✅ directly queryable vs ❌ cited-within-review)chapter_XX/sections/:
→ Call source_registry skill
→ Obtain: author_paper_ids (001), review_ids (002), general_ids
→ Call knowledge skill (retrieve mode)
→ Input: section topic keywords
→ Output: existing insights that can skip or accelerate research
Research is guided by references.md — the literature groundwork done during /chapter N. Start with the sources already identified as relevant to this section, then broaden.
Parse references.md for this section: Read Section 3 ("References Breakdown by Section") to find the key sources listed for this specific section (X.Y or X.Y.Z). From the Reference Data Table (Section 4), extract ✅ Direct Sources (individually queryable in NotebookLM) and ❌ Referenced Sources (cited within reviews — note which review(s) to query from the "Cited In" column). Build a prioritized query plan: these sources are queried first.
Priority queries — section-specific sources:
# For ✅ Direct Sources: query them directly
mcp_notebooklm_notebook_query(
notebook_id="1b7df790-7858-4fc8-879c-39f41238c4ae",
query="<specific question about a key point in section X.Y>",
source_ids=<direct_source_ids_from_references_md>
)
# For ❌ Referenced Sources: query the review(s) that cite them
mcp_notebooklm_notebook_query(
notebook_id="1b7df790-7858-4fc8-879c-39f41238c4ae",
query="What does this review say about [Referenced Paper] regarding [topic]?",
source_ids=<citing_review_source_ids>
)
Broadening queries — fill gaps: After priority queries, identify any key points from the outline that still lack sufficient detail. For these gaps:
002) source IDs): consensus definitions, equations, structural contextsource_ids): specific citations, numbers, arXiv IDsAuthor papers (if applicable): Only if this section references the author's own work (001) source IDs).
Track provenance: For each claim, record which source backs it → \cite{bib_key}. Prefer bib keys already in references.md over new ones.
Before drafting, identify and download figures from the literature that would strengthen this section.
Query NotebookLM for figure recommendations:
mcp_notebooklm_notebook_query(
notebook_id="1b7df790-7858-4fc8-879c-39f41238c4ae",
query="For the topics covered in section [X.Y] about [topic],
which published figures are commonly used to illustrate
these concepts? For each, provide the paper's arXiv ID,
figure number, and a brief description.",
source_ids=<review_002_ids>
)
Cross-reference with references.md Figure Candidates Table (if literature_research Phase 3 was run for this chapter) — merge any pre-identified candidates.
Attempt download for each candidate figure:
# Get figure URLs from InspireHEP
mcp_inspirehep_get_paper_figures(arxiv_id="https://arxiv.org/abs/<ID>")
# Match the target figure number from the returned list
# Download to chapter directory
# run_command: curl -L -o "chapter_XX/figures/<bib_key>_figN.png" "<figure_url>"
See paper_lookup Recipe 3 for full details.
Handle failures: If the figure is not indexed on InspireHEP, record it for a % TODO placeholder in the draft (see Step 6 fallback below).
scientific-writing): Outline key points from Step 4 research, then convert to flowing paragraphshumanizer): Anti-AI pass — strip patterns, vary rhythmOutput format (LaTeX-ready markdown):
$$ for display, $ for inline\cite{bib_key}\label{sec:X.Y.Z}, \ref{sec:...}Literature figures: Insert figures identified in Step 5. For each successfully downloaded figure, use:
\begin{figure}[t]
\centering
\includegraphics[width=\columnwidth]{figures/<bib_key>_figN.png}
\caption{<Description of what the figure shows>, from \cite{<bib_key>}, Figure~N.}
\label{fig:<descriptive_label>}
\end{figure}
If a figure was not available on InspireHEP (Step 5 failure), insert a placeholder with a % TODO comment:
\begin{figure}[t]
\centering
% TODO: insert Figure 3 from arXiv:2302.01947 (Amerio et al., 2023)
\includegraphics[width=\columnwidth]{figures/placeholder.pdf}
\caption{Reconstructed source-count distribution from \cite{Amerio:2023uet}, Figure~3.}
\label{fig:dnds_result}
\end{figure}
The % TODO comment must include: (1) the figure number, (2) the arXiv ID or DOI, and (3) the author/year for easy lookup. The user will replace the placeholder with the actual figure file.
This step runs after every subsubsection draft and is the quality gate that must pass before proceeding.
Do NOT proceed to the next subsubsection until this self-review is complete and any identified issues have been revised in the draft.Announce: "Self-review GATE for X.Y.Z: I will not begin X.Y.(Z+1) until I have critically reviewed X.Y.Z, identified weaknesses, and revised them."
A gate violation is: moving to the next subsubsection with known unresolved issues, skipping any of the 6 review checks below, or failing to revise after identifying problems.
Perform the following checks and revise the draft to fix any issues found:
After revisions, save the updated draft to chapter_XX/sections/X.Y.Z_title.md and proceed to the next subsubsection.
Run this step only after all subsubsections have passed their self-quality gates:
1. Chapter-level coherence:
2. Document-level coherence:
outline.md?Fix any issues found, then proceed to Step 9.
After coherence check passes, invoke the referee skill using the Skill tool automatically on the assembled section:
/referee X.Y)chapter_XX/sections/chapter_XX/sections/X.Y_review.md for the user's referenceThis review pass is automatic — the agent does not wait for user input. It serves as a final quality filter before presenting to the user.
This step runs once, after the review pass is complete.
.tex snippet ready for \input{} in chapter_X.texchapter_XX/sections/X.Y.Z_title.md — draft in LaTeX-ready markdownchapter_XX/sections/X.Y.Z_title.tex — LaTeX conversion (only when requested).agent/knowledge/ — key insights saved via the knowledge skill (if new research was performed)