AutoR

Adversarial but evidence-bound Reviewer

Plan a literature review before drafting. Build a lean evidence-bound, taste-gated package using references.bib, reference-map.json, review-plan.md, evidence-ledger.md, and table-figure-plan.md.

Draft a fact-grounded, taste-gated review manuscript from an approved lean planning package. Uses references.bib, reference-map.json, review-plan.md, evidence-ledger.md, and table-figure-plan.md; use only after plan is approved.

Understand the autor codebase at a high level. Use when the user asks how to use autor, what built-in skills it has, what features it supports, which workflow to choose, or wants a project overview before using other skills.

Write an exploratory literature review based on papers in an autor workspace. For approved review-article workflows, respect the canonical references.bib / reference-map.json / evidence-ledger contract produced by plan.

High-level orchestrator for iterative review planning and hypothesis evolution. Builds the current autor planning package, triggers targeted acquisition, maintains corpus layers, and stops before writing.

Polish academic writing for publication — remove AI and workflow artifacts, sharpen conclusion-led judgment, normalize terminology, improve clarity, and rewrite draft-like or self-referential prose into submission-ready scholarly text. Supports both Chinese and English.

Use the Records-backed AutoDownload service as autor's external literature acquisition service. Choose this skill when the user needs PubMed retrieval, PMID resolution, or PDF download to support an autor workspace.

Verify citations in AI-generated or human-written text against the local knowledge base and external APIs. Catches hallucinated references, wrong metadata, and missing papers. Use when the user wants to check if citations are real and accurate.

Generate and inspect Office documents (DOCX, PPTX, XLSX). Generate by writing Python scripts that call python-docx, python-pptx, and openpyxl APIs directly. Inspect with `autor document inspect` to verify layout, content, and catch issues (overflow, missing elements). Use when the user wants to create Word reports, PowerPoint presentations, Excel data sheets, or inspect any Office document.

Enrich paper metadata using LLM extraction. Extract table of contents (TOC), generate L3 paper-level conclusion cards, backfill abstracts, or refetch citation counts from APIs. Use when the user wants to build TOC, update L3, update citation data, or backfill missing abstracts.

Export papers from the knowledge base to standard citation formats like BibTeX. Supports exporting all papers, specific papers, or filtered by year/journal. Use when the user needs BibTeX entries, reference files, or citation export.

Query citation graphs — view a paper's references, find which papers cite it, and analyze shared references between multiple papers. Use when the user asks about citation relationships, reference overlap, or bibliographic connections.

Import papers from external reference managers (Endnote XML/RIS, Zotero Web API or local SQLite). Handles PDF matching, MinerU conversion, metadata enrichment, and index updates. Use when the user wants to import their existing library from Zotero, Endnote, or attach a PDF to an existing paper.

Rebuild the node-level SQLite FTS5 evidence index. Use when the user asks to rebuild search, refresh search after metadata/full-text changes, or repair missing search results.

Ingest papers from inbox into the knowledge base. Runs the pipeline to convert PDFs via MinerU (auto-splits long PDFs), extract metadata, deduplicate by DOI, generate L3 paper-level conclusion cards by default, and update the node-level FTS5 evidence index. Supports three inboxes - regular papers, theses, and general documents.

Enhance biomedical figure prompts before image generation. Read the source manuscript, infer missing molecules, cells, tissues, and anatomical relationships, and output a long, scientifically constrained English prompt in a BioRender × Nature Reviews style for downstream plotting.

Identify research gaps and open questions from the literature in a workspace. Combines auditable search, citation analysis, and cross-paper comparison. Use when the user wants to find unexplored areas, formulate research questions, or assess where the field is heading.

Draft point-by-point responses to peer review comments. Locates supporting evidence from workspace papers and the original manuscript. Use when the user receives reviewer feedback and needs to write a rebuttal or revision response letter.

Search academic papers in the local autor knowledge base. Uses auditable node-level FTS5 evidence search, evidence bundles, author search, and top-cited ranking. Use when the user wants to find papers, look up literature, search by author, or explore research topics.

Initialize and diagnose the autor environment. Run interactive setup wizard (bilingual EN/ZH) to install dependencies, create config files, and configure API keys. Run status check to see what's installed and what's missing. Use when the user wants to set up, install, configure, or troubleshoot autor.

View paper content at different detail levels. L1 (metadata), L2 (abstract), L3 (paper-level conclusion card), L4 (full text). Use when the user wants to read a paper, see its abstract, L3 takeaway, or full content.

Retrieve structured clinical trial records by disease, intervention, phase, status, location, or a free-text treatment theme, and automatically save the search plan, normalized JSON, and Markdown summary into a workspace. Use when the user wants trial retrieval that stays attached to the current review workflow.

Revise a manuscript under reviewer or editor comments with traceable, fit-for-comment edits. Uses the current autor citation-key contract and compact revision package; preserves unaffected content but escalates to true rewrites when reviewer intent requires it.

Use this skill after a review draft is complete. It launches an independent checking sub-agent near the end of the manuscript to systematically inspect the review's structure, prose quality, section-opening seed quality, mechanistic argumentation, clinical-evidence organization, analytical value of tables, and AI-like tone. If the draft falls short, surface polishing is not allowed; the specified sections must be sent back and rewritten against a concrete issue list.

Audit paper data quality in the knowledge base. Checks for missing fields, filename issues, DOI duplicates, title mismatches, and more. Supports LLM-based deep diagnosis for title mismatches and automated repair. Use when the user wants to check data quality, find problems, or fix metadata issues.

View top-cited papers ranking and refetch citation counts from APIs. Use when the user asks about highly cited papers, citation rankings, or wants to update citation data.

View LLM token usage, API call timing, and runtime metrics. Use when the user asks about token consumption, API costs, or performance statistics.

Rename paper directories to standardized Author-Year-Title format based on JSON metadata. Use when the user wants to normalize filenames after metadata corrections.