Search Skills

Creates a 24-48 hour literature snapshot in bullet-point format from a core paper set, using only provided citations and avoiding long paragraphs.

Creates a peer review report in Markdown format using extracted claims and evidence gaps, covering novelty, soundness, clarity, and impact with actionable feedback.

Automates creation of citation-backed markdown tables from schema and evidence packs, eliminating placeholders and ensuring strict citation adherence.

Generates a TODO report identifying evidence gaps in survey documentation, guiding upstream fixes to improve content quality before proceeding.

Prevents citation drift across markdown subsections after document polishing, ensuring claim-evidence alignment remains intact.

Automates expansion and refinement of search keywords (synonyms, acronyms, exclusions) to improve coverage and reduce noise, updating queries.md configuration.

Automates citation density improvement in technical documents by redistributing existing citations to meet unique citation requirements without adding new facts or changing content.

Creates structured evidence packs with claim candidates, comparisons, and citation-provenance snippets for document sections using existing bibliography.

Converts Markdown drafts into structured LaTeX projects with bibliography wiring, removing markdown formatting and linking to a .bib file.

Injects existing citations from reference files into drafts to satisfy global unique-citation requirements without introducing new facts.

Automates binding of evidence IDs from a pre-existing bank to document subsections (H3), enabling section-scoped evidence usage in technical writing.

Automates merging fragmented document outlines to achieve a target section count (6-8 H2 sections) without generating prose.

Generates chapter lead blocks that connect H3 subsections and preview the chapter's comparison lens without adding new facts or citations.

Drafts non-prose visual artifacts (tables, timelines, figure specs) for surveys using evidence and BibTeX citations.

Detects merge-induced voice inconsistencies in documentation and routes fixes to source files without modifying the draft.

Automates generation and verification of BibTeX entries from paper notes, providing traceable references for prose and LaTeX with URL, date, and title validation.

Creates academic paper front matter sections (Abstract, Introduction, etc.) with high citation density and evidence-policy paragraphs, adhering to provided bibliography.

Automates merging individual section files into a single draft document while preserving outline order and weaving transitions for review and polishing.

Produces coverage and redundancy reports for outline/mapping files, validating structure before content creation with no prose output.

Creates structured academic notes (summary/method/results/limitations) for papers in a core set, stored as JSONL for evidence-based writing.

Generates multi-level taxonomies from research paper sets with descriptive node mappings for clear topic organization.

Processes raw paper data to produce a deduplicated set and a ranked core set for analysis and organization.

Generates evidence-based novelty matrices comparing submission contributions against related work for academic peer review documentation.

Provides AI-powered code reviews with Karen Scores (0-100) assessing over-engineering, honesty, and practical value, available as GitHub Action or IDE tool.