菜单
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.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
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.
| name | audit |
| description | 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. |
Check the data quality of papers already in the knowledge base. Runs in stages: rule-based checks (automatic) → LLM deep diagnosis (on suspicious items) → automated repair.
autor audit [--severity error|warning|info]
Issues are classified by severity:
For each title_mismatch paper, use the Read tool to load meta.json and the first 80 lines of paper.md, then determine:
For confirmed mismatches, use the repair command:
# Preview with dry-run first
autor repair "<paper-id>" --title "Correct Title" [--author "First Author"] [--year YYYY] [--doi "10.xxx/..."] --dry-run
# Apply after confirmation
autor repair "<paper-id>" --title "Correct Title" [--author "First Author"] [--year YYYY] [--doi "10.xxx/..."]
# Rebuild the index after repair
autor pipeline reindex
| Rule | Level | Description |
|---|---|---|
missing_title | error | Title is absent |
missing_md | error | No MD file corresponding to the JSON |
duplicate_doi | error | Duplicate DOI |
missing_doi | warning | DOI is absent |
missing_abstract | warning | Abstract is absent |
title_mismatch | warning | JSON title does not match MD H1 |
nonstandard_filename | info | Filename does not follow the standard format |
User says: "Check whether there are any problems in my paper library." → Run Stage 1 rule-based check
User says: "Run a deep check." → Run Stage 1 + Stage 2 (LLM diagnosis on each title_mismatch paper)
User says: "Fix the mismatched papers." → Run Stage 3