Thesis production orchestrator
Skill lihongqing2001-gif/chinese-thesis-production-skills/skills/thesis-production-orchestrator
Evidence-first Codex skills for guided Chinese STEM thesis production
npx -y skills add lihongqing2001-gif/chinese-thesis-production-skills --skill thesis-production-orchestratorAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 13 days oldThe repository was created 13 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Guide and orchestrate an evidence-first Chinese STEM degree-thesis project from first-run intake through submission. Use when a new Codex user starts, resumes, restructures, plots, formats, revises, recovers, or audits a thesis; when the authoritative manuscript is unclear; or when literature, methods, data analysis, Origin plots, chapters, figures, citations, Word, reviews, and optional computation must be coordinated.
SKILL.md
5.0 KB, as published. Nobody here has run it
Thesis Production Orchestrator
Manage the thesis as a versioned evidence system. Do not begin by rewriting prose.
First-run sequence
- Determine whether the user is starting, restructuring, writing one chapter, revising, formatting, plotting, auditing, or recovering.
- Run the intake questionnaire in references/onboarding.md.
- Generate the answers JSON and execute
guided_intake.py; do not require the user to construct manifests by hand. - Route school rules, outline, scope, and chapter contracts to
thesis-requirements-planning. - Ask the author to identify the authoritative manuscript when it cannot be established from a written decision. Never infer authority from
final, version numbers, timestamps, page count, or folder location. - Record SHA-256 hashes before any edit and complete material, conflict, gap, and terminology ledgers.
- Use Superpowers
brainstormingfor scope/design,writing-plansfor multi-stage work, andverification-before-completionbefore any completion claim. - Route work to the narrowest specialist Skill and run its deterministic gate.
- Deliver new files plus a report of inputs, changes, validation, unresolved items, and required human approvals.
Routing
- School rules, thesis structure, supervisor framework, scope, and staged plan: use
thesis-requirements-planning. - Review search, field mapping, authoritative reviews, primary studies, publication figures, or ChatGPT web research handoffs: use
thesis-literature-review. - Zotero, references, DOI checks, evidence matrix, experimental facts, or factual support: use
thesis-evidence-engineering. - Materials, preparation, characterization conditions, controls, replicates, statistics, or method-result coverage: use
thesis-research-methods. - Raw data, OriginPro, Python/R plots, fitting, error bars, graph templates, or plot specifications: use
thesis-data-analysis-plotting. - Chapter outline, research logic, section integration, or prose: use
thesis-chapter-engineering. - Figures, panels, captions, tables, or visual readability: use
thesis-figure-table-audit. - DOCX, school template, page layout, equations, TOC, headers, or references layout: use
thesis-docx-production. - Reviewer, defense, supervisor, or blind-review comments: use
thesis-review-revision. - CFD, MD, finite-element, numerical models, source terms, validation, or simulation figures: optionally use
thesis-computational-evidenceonly when the research requires it.
Load only the selected specialist Skill and relevant reference files.
Authority rules
Apply the hierarchy in references/versioning.md. A lower-ranked artifact may reveal a conflict but may not silently override a higher-ranked source.
Treat these as separate categories:
- verified literature fact;
- author-provided experimental fact;
- completed computational result;
- inference consistent with evidence;
- recommendation or future work.
Do not merge the categories in prose or reports.
Edit modes
Choose one mode before editing:
draft: create new Markdown or DOCX from controlled inputs.restructure: change chapter organization while preserving registered evidence and assets.content-patch: modify only declared paragraphs or sections.format-patch: preserve text and modify only Word structure/style.review-pack: leave the manuscript untouched and produce paste-ready revisions.recovery: compare versions, restore lost text/figures, and establish a new baseline.
Never combine restructure, content-patch, and format-patch without an explicit change scope.
Deterministic tools
Initialize a project:
python scripts/init_thesis_project.py /path/to/project --title "Title"
Preferred guided initialization:
python scripts/guided_intake.py /path/to/project --answers intake-answers.json
Record a baseline:
python scripts/snapshot_project.py /path/to/project \
--output /path/to/project/versions/baseline.json
Run a stage gate:
python scripts/quality_gate.py /path/to/project --stage evidence \
--report /path/to/project/reports/evidence-gate.json
Read references/project-schema.md before changing artifact fields. Read references/gates.md before declaring a stage complete.
Completion rule
Do not call the thesis or a stage complete because a document was generated. Require fresh audit output, visual inspection where applicable, unchanged baseline verification, and an explicit list of unresolved human decisions.