Skip to content
CAI
Software that uses CAICheck a score

MiniMax-AI/skills

46.0

Weak · 11 November 2026

17.2k

lines of production code

C#

with Python

7

measurements over time

CAI band scale
CAI trend line
CAI lens gauges

What this system is

Features

Automated PR validation for skill submissions

A new automated validation script and associated review guidelines have been added to enforce structural and quality standards for new skill submissions. The \validate\_skills.py\ script checks that every skill directory contains a valid \SKILL.md\ with required YAML frontmatter fields (\name\, \description\), ensures the name matches the directory, and scans for hardcoded secrets (OpenAI keys, AWS keys, Bearer tokens). This automation is paired with a two-phase PR review process: Phase 1 runs the script to block structural errors, while Phase 2 applies manual quality guidelines covering skill scope, description clarity, file size, credential handling, script quality, language, and README synchronization.

.claude · high confidence

Introduce minimax-pdf skill for polished PDF generation, form filling, and document reformatting

A new 'minimax-pdf' skill has been added to the skills directory, providing three distinct capabilities for PDF handling. First, it allows users to CREATE visually polished PDFs from scratch using a token-based design system that applies specific color palettes, typography, and cover patterns (such as 'fullbleed', 'split', or 'typographic') based on document types like 'report', 'resume', or 'proposal'. Second, it enables users to FILL form fields in existing PDFs by inspecting and writing values to text, checkbox, dropdown, and radio fields without altering the layout. Third, it supports REFORMATting existing documents (Markdown, text, or PDF) by parsing them into a content structure and applying the same design system to generate a styled PDF output. The skill relies on Python libraries (reportlab, pypdf) and Node.js (Playwright) for rendering.

skills/minimax-pdf · high confidence

Introduce minimax-xlsx skill for Excel file operations

Added the minimax-xlsx skill, providing a comprehensive guide and supporting scripts for opening, creating, reading, analyzing, editing, and validating Excel files (.xlsx, .xlsm, .csv, .tsv). The skill enforces a strict XML-based editing workflow to preserve advanced features like VBA, pivot tables, and charts, which are typically lost when using standard libraries like openpyxl. It includes utilities for formula validation, financial formatting standards, and safe row/column manipulation.

skills/minimax-xlsx · high confidence

New GIF Sticker Maker skill for creating animated Funko Pop-style stickers

Users can now convert photos of people, pets, objects, or logos into four animated GIF stickers in a Funko Pop / Pop Mart 3D figurine style. This new skill in the \skills/gif-sticker-maker\ directory orchestrates a workflow using the MiniMax Image and Video Generation APIs: it generates static sticker images, animates them into short videos, and converts the results to GIFs using ffmpeg. The skill supports character reference images to preserve facial likeness for person subjects, handles multilingual captions, and outputs the final assets via a structured delivery block.

skills/gif-sticker-maker · high confidence

New frontend-dev skill with integrated asset generation and licensed fonts

The \skills/frontend-dev\ directory now includes a comprehensive skill definition (\SKILL.md\) that guides the creation of production-ready frontend pages with cinematic animations, AI-generated media, and persuasive copy. This update adds Python scripts for generating images, video, audio, and text-to-speech assets via Minimax, along with a curated library of SIL Open Font License fonts (such as Arsenal, Big Shoulders, and Geist Mono) in the \canvas-fonts\ directory to ensure compliant visual design without relying on external placeholder services.

skills/frontend-dev · high confidence

New minimax-docx skill for professional DOCX creation and formatting

The new minimax-docx skill enables the creation, editing, and formatting of Word documents using the OpenXML SDK. It provides three main workflows: creating new documents from scratch, filling or editing content in existing files, and applying template styles with XSD validation. The skill includes pre-defined academic, corporate, and default style sets, along with XSD schemas for aesthetic and business rule validation to ensure document quality and compliance.

skills/minimax-docx · high confidence

Dependencies

Add gif-sticker-maker and minimax-docx skills with their dependencies

This change introduces two new skills: gif-sticker-maker, which adds a Python-based tool requiring the requests library (\>=2.28) and a system-installed ffmpeg, and minimax-docx, which adds a .NET CLI and core library targeting net8.0. The minimax-docx skill depends on DocumentFormat.OpenXml (3.5.1) and System.CommandLine (2.0.5) to handle document processing.

(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 44 → 46 (+1.6)
  • Rubric changed (rubric-2026.10.5 → rubric-2026.11.2) — scores are not directly comparable.

Lenses

  • Code Health 86 → 84 (-1.9)
  • Architecture 97 → 97 (+0.0)
  • Maturity 61 → 68 (+7.0)
  • Readiness 10 → 10 (+0.0)
  • Security 89 → 89 (+0.0)

Resolved (11)

  • Duplicated block (10 lines × 2) (skills/minimax-pdf/scripts/render_body.py)
  • Duplicated block (11 lines × 2) (skills/frontend-dev/scripts/minimax_image.py)
  • Duplicated block (16 lines × 2) (skills/frontend-dev/scripts/minimax_image.py)
  • Duplicated block (5 lines × 3) (skills/minimax-xlsx/scripts/xlsx_add_column.py)
  • Duplicated block (6 lines × 2) (skills/frontend-dev/scripts/minimax_image.py)
  • Duplicated block (7 lines × 4) (skills/frontend-dev/scripts/minimax_image.py)
  • Duplicated block (9 lines × 2) (skills/frontend-dev/scripts/minimax_image.py)
  • Members sharing a duplicated core (4 members, 50+ identical tokens) (skills/frontend-dev/scripts/minimax_image.py)
  • Near-duplicate member pair (26 shared lines) (skills/frontend-dev/scripts/minimax_image.py)
  • No tests found
  • XML-doc coverage: MiniMaxAIDocx.Core (skills/minimax-docx/scripts/dotnet/MiniMaxAIDocx.Core/MiniMaxAIDocx.Core.csproj)

New (9)

  • Duplicated block (7 lines × 3) (skills/frontend-dev/scripts/minimax_music.py)
  • Edited copy of a member (14 corresponding lines) (skills/minimax-pdf/scripts/cover.py)
  • Edited copy of a member (22 corresponding lines) (skills/frontend-dev/scripts/minimax_video.py)
  • Edited copy of a member (25 corresponding lines) (skills/minimax-docx/scripts/dotnet/MiniMaxAIDocx.Core/Samples/FieldAndTocSamples.cs)
  • Edited copy of a member (48 corresponding lines) (skills/minimax-docx/scripts/dotnet/MiniMaxAIDocx.Core/Samples/AestheticRecipeSamples.cs)
  • minimax_image.main (cognitive 21) (skills/frontend-dev/scripts/minimax_image.py)
  • redundant comment (skills/minimax-docx/scripts/dotnet/MiniMaxAIDocx.Core/Commands/CreateCommand.cs)
  • redundant comment (skills/minimax-docx/scripts/dotnet/MiniMaxAIDocx.Core/Commands/CreateCommand.cs)
  • redundant comment (skills/minimax-docx/scripts/dotnet/MiniMaxAIDocx.Core/OpenXml/RunMerger.cs)

Changes since last survey

  • 4 commits — 2 feature/other, 2 fixes

By area

  • (repo) — 2 commits
  • skills/frontend-dev — 1 commit
  • skills/minimax-multimodal-toolkit — 1 commit

Notable commits

  • fix: Merge pull request #128 from SaladDay/fix/frontend-dev-h3-image
  • fix: fix(frontend-dev): switch image script to MiniMax-H3-Image
  • change: Merge pull request #127 from SaladDay/docs/minimax-multimodal-toolkit-h3
  • change: docs(minimax-multimodal-toolkit): sync mmx-cli skill with v1.0.29 (H3-Image, M3.1-Flash)

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

Survey your own repository

MiniMax-AI/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 11 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 02853d885acb1abc157477ce00c4d98a8ed14a71 — the exact code this score is about.
  • Scored under rubric-2026.11.2 — 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-2026.11.29+preprod-106b68c3c303.