Skip to content
CAI
Software that uses CAICheck a score

Yuan1z0825/nature-skills

55.0

Adequate · 15 November 2026

31.8k

lines of production code

Python

with JavaScript

5

measurements over time

CAI band scale
CAI trend line
CAI lens gauges

What this system is

Features

Add figures4papers demo assets and third-party notices

The nature-figure skill now includes a collection of reference scripts and preview images from the upstream figures4papers project. This addition introduces a THIRD\_PARTY\_NOTICES.md file to clarify that these assets are retained for source attribution and research reference only, and do not inherit the repository's root MIT License. The new assets include Python plotting scripts for generating scientific figures, such as ablation studies for CellSpliceNet, cross-species method comparisons, diffusion swiss roll visualizations, gene regulatory network analyses, and frequency prior plots, providing users with reference implementations for generating high-quality research figures.

skills/nature-figure/assets · high confidence

Add initial submission LaTeX templates

Added four new LaTeX templates for manuscript submission components: a title page with fields for author details, affiliations, and required declarations; an initial cover letter structured for journal submission; a document for submission declarations and reviewer suggestions (including tables for suggested and opposed reviewers); and a highlights and editorial summary template. All templates use AUTHOR\_INPUT\_NEEDED placeholders to guide authors in providing specific manuscript details.

skills/nature-writing/templates · high confidence

Add no-MCP fallback academic search and citation conversion scripts

The nature-academic-search skill now includes a standalone, no-MCP fallback search capability in \academic\_search.py\ that queries the OpenAlex API for papers by topic, author, ORCID, or institution, and provides a \format-converter.py\ tool to download and convert citations from PubMed, CrossRef, and arXiv into RIS, BibTeX, or ENW formats. This enables the skill to function in environments without the MCP server, such as plain CLI usage or CI pipelines, and includes a \preflight.py\ utility to verify connectivity to the required API endpoints. Also in this area: 1 commit (1 fix) fixing skills.

skills/nature-academic-search/scripts · medium confidence · unverified

Add script to convert Markdown proposals to formatted DOCX

A new Python script, \build\_proposal\_docx.py\, has been added to the \skills/nature-proposal-writer/scripts\ directory. This tool converts research proposal Markdown files into professionally formatted Microsoft Word documents, applying specific academic styling standards such as Times New Roman and SimSun fonts, 1.5 line spacing, 1-inch margins, and proper table grid styles.

skills/nature-proposal-writer/scripts · high confidence

Added templates for revision correspondence and manuscript markup

New LaTeX templates have been added to support the revision process for journal submissions. These include a cover letter template for addressing the editor, a structured response-to-reviewers template that enforces one-to-one reviewer isolation and point-by-point responses, and a redline manuscript template for marking changes in the revised text.

skills/nature-response/templates · high confidence

Documentation for academic search reference handling and workflows

Added comprehensive reference documentation for the nature-academic-search skill, including strategies for parsing citations from various document formats (.docx, .tex, .txt), a unified deduplication engine for merging literature results, source tier classifications for API-backed and scraped data sources, and detailed workflow definitions for multi-source search, citation verification, MeSH strategy, citation file management, and reference management. Also included format specifications for RIS, BibTeX, and EndNote exports, and a new strict other-citation impact audit workflow for analyzing citation independence and impact.

skills/nature-academic-search/references · high confidence

Introduce Image2PPT skill with comprehensive QA and validation scripts

This change adds the \nature-image2ppt\ skill, providing a suite of Python scripts to manage the conversion of images to PowerPoint presentations. The skill includes \postprocess\_manifest\_arrows.py\ to apply arrow extensions to PPTX slide XML, \inspect\_arrow\_atomicity.py\ and \inspect\_region\_decomposition.py\ to validate that manifest arrows map correctly to PowerPoint objects and that region decompositions match standard manifest objects, and \render\_image2ppt\_qa.py\ to render slides via PowerPoint COM or LibreOffice for visual comparison. The \run\_image2ppt\_qa.py\ and \run\_final\_image2ppt\_qa.py\ scripts orchestrate these checks as page-level and final deck gates, while \visual\_review\_evidence.py\ generates and validates structured review evidence. \runtime\_paths.py\ ensures all file operations are confined to the skill's directory to prevent path traversal.

skills/nature-image2ppt/scripts · high confidence

Introduce academic search MCP server with multi-source support

Adds a new MCP server for the nature-academic-search skill that provides a unified search tool across CrossRef, PubMed, and arXiv, along with optional, explicitly-enabled providers for Scopus and ScienceDirect. The server exposes tools for searching papers, retrieving details by ID, formatting citations, and querying Scopus metadata and metrics, with configuration managed via environment variables and a local config file.

skills/nature-academic-search/mcp-server · high confidence

Introduce academic search data source modules

Added the source modules for the new nature-academic-search skill, providing data access to CrossRef, PubMed, arXiv, Scopus, and ScienceDirect. The implementation includes an arXiv source that uses defusedxml to safely parse Atom feeds, a PubMed source that enforces NCBI rate limits and requires an email configuration, and Elsevier-backed sources (Scopus and ScienceDirect) that initialize the pybliometrics library from a config file and convert records to JSON-safe structures.

skills/nature-academic-search/mcp-server/sources · high confidence

Introduce nature-academic-search skill with router-based workflow architecture

The \nature-academic-search\ skill is now available, enabling multi-source scholarly search (CrossRef, PubMed, arXiv, Scopus, ScienceDirect), citation formatting, and strict external-citation audits. It uses a static/dynamic router architecture defined in \manifest.yaml\ and \SKILL.md\ to map user requests to specific workflows (e.g., multi-source search, citation verification, MeSH strategy, citation-file management, reference management, and strict other-citation impact audit). The skill includes an installer (\install.sh\) that sets up the MCP server, configures \.mcp.json\, and enables the skill in Claude Code settings. It supports both MCP-based and no-MCP fallback search (via \scripts/academic\_search.py\) and requires local configuration for PubMed and Elsevier sources.

skills/nature-academic-search · high confidence

Introduce nature-downloader skill for lawful academic full-text retrieval

The new nature-downloader skill provides a structured, auditable way to obtain academic full text through lawful open-access routes, publisher APIs (Elsevier, Springer Nature, IEEE), CNKI institutional access, or the user's own institution-authorized browser session. It enforces explicit Supporting Information (SI) selection before any download, routes requests based on language and publisher, and falls back to OA sources or Web Access only when primary routes fail. The skill ships with a Node.js batch download entry point, Python configuration helpers for school and publisher credentials, JSON/YAML schemas for school configuration, publisher routing data, and detailed documentation covering first-run setup, credential handling, and safety boundaries.

skills/nature-downloader · high confidence

Introduce nature-paper-trans skill for layout-preserving English-to-Chinese PDF translation

Adds a new skill that translates English research-paper PDFs (full text or selected pages) into Simplified Chinese image-based PDFs while preserving original page dimensions, orientation, and layout. The workflow splits the source PDF, generates one translated image per page using the built-in image\_gen tool, requires a manual review record for each page before assembly, and merges the results into a single Chinese PDF. It supports up to 10 concurrent page requests, allows resuming interrupted tasks, and lets users update specific pages; it does not produce editable text or OCR layers.

skills/nature-paper-trans · high confidence

Introduce nature-shared shared reference support package

A new \nature-shared\ support package has been added to centralize definitions and reference materials used by multiple \nature-\*\ skills (such as \nature-writing\ and \nature-polishing\). This package includes shared core guidelines for manuscript consistency, discussion argumentation, main-text discipline, and section-specific writing patterns (Abstract, Introduction, Results/Discussion), along with journal format rules for flagship Nature, Nature Communications, and Nature Machine Intelligence. It also provides a Python-based consistency scanning script to detect terminology variants and numeric inconsistencies, serving as a mechanical first-pass audit for users rather than an automated editor.

skills/nature-shared · high confidence

Introduce the nature-image2ppt skill for converting images and PDFs to editable PowerPoint decks

This change adds the \nature-image2ppt\ skill, providing a CLI tool (\image2ppt\) that converts image files and PDFs into editable PowerPoint presentations. The runtime handles input normalization (including flattening transparent images onto a white background), extracts text hints via PaddleOCR or a built-in offline detector, renders LaTeX formulas, and supports image generation/editing through a provider-neutral backend (Codex OAuth or OpenAI-compatible API). It manages deck run state, validates page readiness, and builds the final PPTX with content-box positioning and slide-size constraints.

skills/nature-image2ppt/cli · high confidence

Introduces automated verification handling and CNKI support to the nature-downloader skill

The nature-downloader skill now includes a new \anti-bot.mjs\ module that automatically attempts to pass common publisher verification challenges (such as simple click prompts, ScienceDirect robot checks, Cloudflare Turnstile, and CNKI slider CAPTCHAs) via the user's authenticated Chrome session, falling back to user intervention only when necessary. Additionally, the skill adds dedicated support for the CNKI database, including a new \cnki.mjs\ library for search and download routing, and updates the \credentials.mjs\ and \provider-registry.mjs\ to manage provider-specific secrets and registry data securely.

skills/nature-downloader/scripts · high confidence

New automation and validation scripts for skills management and documentation

This change introduces a suite of new shell and Python scripts in the \scripts/\ directory to automate skill updates, validate repository integrity, and enhance documentation. \autoupdate-skills.sh\ enables automatic, throttled synchronization of installed skills from upstream, while \update-codex-skills.sh\ handles manual installation and pruning of skills into the Codex environment. Documentation and metadata integrity are enforced by new validators: \validate-skill-metadata.py\ checks skill frontmatter and interface contracts, \validate-skill-index.py\ ensures the README skill list matches the repository, \validate-readmes.py\ and \validate-readme-mirror.py\ verify README consistency and bilingual mirroring, and \validate-workflows.py\ ensures CI path filters and action pinning are correct. Additionally, \generate-star-history.py\ creates static star-history SVGs, and \update-reference-tocs.py\ adds navigation lists to long reference files.

scripts · high confidence

New nature-figure scripts for panel alignment, safety, and schematic generation

The nature-figure skill now includes a suite of new scripts to improve figure quality and generation. A new panel alignment audit system (audit\_panel\_alignment.py and panel\_alignment.R) validates multi-panel figure layouts at render time to ensure visual consistency. Figure safety helpers (figure\_safety.py) prevent silent numerical errors in interpolation and label positioning. A new script (generate\_openrouter\_schematic.py) allows generating manuscript schematics via OpenRouter, with strict security checks to prevent downloading from local or private hosts. Additionally, a backend configuration script (nature\_figure\_backend.py) and updated plot templates (plot\_templates.py) provide better support for Python and R workflows.

skills/nature-figure/scripts · high confidence

New nature-paper-card skill for structured scientific paper analysis

A new \nature-paper-card\ skill has been added to the \skills/nature-paper-card\ directory, enabling the generation of a structured, source-grounded 16-section Paper Card for deep-reading single scientific papers. This skill focuses on analyzing research questions, method logic, experiment-to-claim evidence chains, conclusion boundaries, and generating testable research ideas, rather than producing simple summaries or translations. It includes a router (\SKILL.md\), a fixed output schema (\references/card-schema.md\), and provenance rules (\references/evidence-and-provenance.md\) to ensure claims are backed by specific evidence. The skill supports three locator modes (\page-grounded\, \structure-grounded\, \source-limited\) to handle varying levels of source reliability, and includes bundled Python scripts (\prepare\_paper.py\, \audit\_paper\_card.py\) for source preparation and quality assurance. It is designed to work alongside other skills like \nature-reader\ and \nature-reviewer\ for broader literature workflows.

skills/nature-paper-card · high confidence

New patent disclosure workflow and conversion scripts

Adds the \nature-paper-to-patent\ skill, providing a complete workflow to convert technical papers and presentations into patent application drafts. This includes scripts to extract text from PDFs (\extract\_pdf\_text.py\), convert Word and PowerPoint files to Markdown (\docx\_to\_md.py\, \pptx\_to\_md.py\), and render mathematical formulas and diagrams as images (\math\_render.py\, \mermaid\_render.py\). The suite also features a \build\_patent\_package.py\ script to assemble the final patent package, a CNIPA patent database crawler for prior art search (\cnipa\_epub\_crawler.py\, \cnipa\_epub\_search.py\), and utilities to convert LaTeX to Word Office Math (\math\_to\_omml.py\) and log revision iterations (\iteration\_dialog\_log.py\).

skills/nature-paper-to-patent/scripts · high confidence

New repository validation and skill-specific quality assurance scripts

Added a suite of Python validation scripts to enforce quality and consistency across the repository and individual skills. The new \scripts/validate-repository.py\ tool checks for missing skill metadata, broken README links, and malformed JSON/TOML configs. Skill-specific additions include \nature-citation\'s RIS export and author metadata preservation, \nature-figure\'s PDF collision and font-size auditing, \nature-paper-to-patent\'s Chinese patent claim and draft validation, \nature-paper2ppt\'s PPTX quality checks, and \nature-reader\'s equation rendering validation.

python · high confidence

Behavioural changes

1 commit (0 fixes) modifying RDA-results

A change to existing behaviour in RDA-results — 1 commit, 7 files: “Add existing RDA figures and privacy-screened statistical results (\#278)” (refs \#278).

RDA-results · medium confidence · unverified

2 commits (1 fix) modifying skills/nature-paper2word

A change to existing behaviour in skills/nature-paper2word — 2 commits (1 fix), 10 files: “fix(nature-paper2word): extract reference formatting from paragraph samples (\#285)”; “feat(nature-paper2word): add Markdown → Word skill (documentation only)” (refs \#285).

skills/nature-paper2word · medium confidence · unverified

Repository license changed from MIT to Apache 2.0

The project's licensing terms have been updated from the MIT License to the Apache License, Version 2.0. This change affects the legal terms under which the software and its documentation can be used, reproduced, and distributed.

(repo-wide) · high confidence

Test coverage

4 commits adding/updating tests in skills/nature-citation/tests; Added automated tests for skill update safety and figure generation audits; Added test coverage for academic search server; Added test fixtures and contract verification for the nature-reviewer skill; Added test suite for nature-response skill behavior and consistency; Added test suite for the nature-image2ppt skill; Added tests for Markdown table parsing and source file preservation; Added tests for figure safety validation and URL download security; Added tests for star history generation script; Added unit tests for the nature-paper-to-patent skill.

Dependencies

Pin CI dependencies and define skill-specific requirements

A pinned requirements file for the GitHub Actions CI environment (.github/requirements-ci.txt) has been added to ensure reproducible test runs, alongside new requirements files for several skills (nature-academic-search, nature-downloader, nature-figure, nature-image2ppt, nature-paper-to-patent) and a Node.js lockfile for the patent disclosure tools. These changes establish explicit dependency versions for the CI environment and individual skill runtimes, replacing or supplementing previous loose dependency specifications.

(dependencies) · high confidence

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

How this codebase got here

Score

  • CAI 53 → 55 (+1.8)
  • Rubric changed (rubric-2026.11.1 → rubric-2026.11.4) — scores are not directly comparable.

Lenses

  • Code Health 48 → 47 (-1.3)
  • Architecture 97 → 97 (+0.0)
  • Maturity 82 → 84 (+1.9)
  • Readiness 43 → 43 (+0.0)
  • Security 79 → 93 (+14.4)
  • Accessibility 61 → 70 (+9.6)
  • Performance 100 → 100 (+0.0)

Resolved (34)

  • Duplicated block (10 lines × 2) (skills/nature-academic-search/scripts/converters.py)
  • Duplicated block (11 lines × 2) (skills/nature-paper-to-patent/scripts/math_to_omml.py)
  • Duplicated block (15 lines × 2) (skills/nature-figure/scripts/validate_figure.py)
  • Duplicated block (15 lines × 2) (skills/nature-image2ppt/cli/image2ppt/runtime/_input_normalization.py)
  • Duplicated block (15 lines × 2) (skills/nature-image2ppt/scripts/inspect_region_decomposition.py)
  • Duplicated block (2–6 lines × 4) (skills/nature-figure/assets/figures4papers/figure_Dispersion/plot_idea.py)
  • Duplicated block (5 lines × 2) (skills/nature-paper-to-patent/scripts/validate_patent_draft.py)
  • Duplicated block (6 lines × 2) (skills/nature-citation/scripts/nature_citation.py)
  • Duplicated block (6 lines × 2) (skills/nature-downloader/src/wizard.py)
  • Duplicated block (6 lines × 2) (skills/nature-figure/scripts/plot_templates.py)
  • Duplicated block (6 lines × 2) (skills/nature-image2ppt/cli/image2ppt/runtime/record_page_result.py)
  • Duplicated block (6 lines × 4) (skills/nature-academic-search/mcp-server/academic_search_server.py)
  • Duplicated block (7 lines × 2) (skills/nature-figure/scripts/plot_templates.py)
  • Duplicated block (8 lines × 2) (skills/nature-academic-search/scripts/converters.py)
  • Duplicated block (8 lines × 2) (skills/nature-citation/scripts/nature_citation.py)
  • Duplicated block (8 lines × 3) (skills/nature-figure/scripts/validate_figure.py)
  • Members sharing a duplicated core (4 members, 50+ identical tokens) (skills/nature-academic-search/mcp-server/academic_search_server.py)
  • Off-boarding risk: anonymized user #1
  • Off-boarding risk: anonymized user #2
  • generate-star-history.fetch_stargazers (cognitive 41) (scripts/generate-star-history.py)
  • …and 14 more

New (32)

  • Edited copies of one member (3 members, 6+ corresponding lines) (skills/nature-image2ppt/cli/image2ppt/runtime/_input_normalization.py)
  • Edited copies of one member (3 members, 9+ corresponding lines) (skills/nature-academic-search/mcp-server/academic_search_server.py)
  • Edited copies of one member (5 members, 2+ corresponding lines) (skills/nature-image2ppt/cli/image2ppt/runtime/deck_run_state.py)
  • Edited copies of one member (7 members, 9+ corresponding lines) (skills/nature-academic-search/mcp-server/academic_search_server.py)
  • Edited copy of a member (10 corresponding lines) (skills/nature-downloader/scripts/extract_pdf_text.py)
  • Edited copy of a member (10 corresponding lines) (skills/nature-image2ppt/cli/image2ppt/runtime/deck_run_state.py)
  • Edited copy of a member (11 corresponding lines) (skills/nature-figure/scripts/audit_pdf_text.py)
  • Edited copy of a member (11 corresponding lines) (skills/nature-figure/scripts/generate_openrouter_schematic.py)
  • Edited copy of a member (11 corresponding lines) (skills/nature-image2ppt/cli/image2ppt/runtime/image_gen.py)
  • Edited copy of a member (13 corresponding lines) (skills/nature-image2ppt/cli/image2ppt/runtime/build_pptx_from_manifest.py)
  • Edited copy of a member (17 corresponding lines) (scripts/validate-repository.py)
  • Edited copy of a member (172 corresponding lines) (skills/nature-image2ppt/scripts/run_final_image2ppt_qa.py)
  • Edited copy of a member (18 corresponding lines) (scripts/validate-workflows.py)
  • Edited copy of a member (2 corresponding lines) (skills/nature-image2ppt/cli/image2ppt/runtime/_page_artifacts.py)
  • Edited copy of a member (21 corresponding lines) (skills/nature-paper-to-patent/scripts/disclosure/md_to_docx.py)
  • Edited copy of a member (3 corresponding lines) (skills/nature-academic-search/scripts/converters.py)
  • Edited copy of a member (31 corresponding lines) (skills/nature-shared/scripts/check_consistency.py)
  • Edited copy of a member (35 corresponding lines) (skills/nature-downloader/src/validators.py)
  • Edited copy of a member (43 corresponding lines) (skills/nature-academic-search/mcp-server/academic_search_server.py)
  • Edited copy of a member (46 corresponding lines) (skills/nature-academic-search/scripts/converters.py)
  • …and 12 more

Changes since last survey

  • 6 commits — 5 feature/other, 1 fixes

By area

  • (root) — 4 commits
  • RDA-results/RDA_shading.pdf — 1 commit
  • skills/nature-paper2word — 1 commit

Notable commits

  • fix: fix(nature-paper2word): extract reference formatting from paragraph samples (#285)
  • change: Add existing RDA figures and privacy-screened statistical results (#278)
  • change: chore: update star history
  • change: chore: update star history
  • change: chore: update star history
  • change: chore: update star history

Architecture

  • Unchanged — 0 containers · 1 contexts · 0 edges

Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

Survey your own repository

Yuan1z0825/nature-skills was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.

About this page

  • The score is its most recent published measurement, taken on 15 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
  • Measured at commit 7bff1d0539c0dfad6c30178bb93ce6dd8a5cebf8 — the exact code this score is about.
  • Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
  • Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.73+preprod-52de90b74216.