SenseNova-Skills - A modular AI office skills library open-sourced by the SenseTime team. - AiBoss

What is SenseNova-Skills?

SenseNova-Skills is an open-source, modular AI office skills library developed by the OpenSenseNova team, designed specifically for Agent runtime. The tool breaks down four key office skills—infographic generation, PPT creation, Excel data analysis, and in-depth research—into independent skills that can be directly integrated into Agent frameworks such as OpenClaw and hermes-agent, enabling AI assistants to possess end-to-end professional office execution capabilities.

Main functions of SenseNova-Skills

Infographic generation:sn-infographic Supports 87 layouts and 66 styles, and transforms complex data and long reports into high-density visual infographics through VLM quality checks and multiple iterations;sn-image-imitate New diagrams can be generated based on the style of the reference diagram;sn-image-resume Automatically generate beautiful resume images.
PPT generation:sn-ppt-entry As a unified entry point, after parsing PDF/DOCX/MD/TXT files, they are distributed to either Creative Mode (16:9 PNG per page) or Standard Mode (Style Specification → Outline → Material Planning → VLM Quality Inspection → HTML → Review → Export PPTX).
Data Analysis:sn-da-excel-workflow Supports multi-table reading, automatic conversion of large files (≥10,000 rows) to Parquet, cleaning, filtering, and cross-table aggregation;sn-da-image-caption Achieve table OCR and chart understanding;sn-da-large-file-analysis Large files are processed using streaming.
In-depth research:sn-deep-research Unified scheduling and planning → Multi-dimensional evidence collection → Comprehensive judgment → Final report, supporting breakpoint resume;sn-research-planning Automatically generate research plans;sn-dimension-research Perform search and cross-validation by dimension;sn-research-synthesis Integrate multi-dimensional reports.
Vertical searchIt covers academic (ArXiv, PubMed), developer (GitHub, HuggingFace), Chinese social media (Bilibili, Zhihu, Douyin) and English social media (Reddit, X, YouTube).

The technical principles of SenseNova-Skills

Layered Skills Architecture: Use Tier 0 (base layer, such as sn-image-base Provides text-based image/image recognition) + Tier 1 (application layer, such as...) sn-infographicThe layered design allows lower-level capabilities to be reused and combined with higher-level skills.
Agent Skills SpecificationEach skill has its own independent directory, accessible through... SKILL.md Declare triggers, capabilities, and execution flows, compatible with OpenClaw and hermes-agent runtimes.
VLM Quality Closed LoopIntroducing a visual language model into the generation of infographics and PPTs allows for multiple rounds of quality checks and scoring to ensure that the output meets publishing standards.
Large file streaming processingFor Excel files with tens of thousands of rows, the following approach is adopted: openpyxl Read-only mode + iter_rows Streaming reads, combined with Parquet transformation and block processing, reduce memory usage.
Resume execution mechanismIn-depth research processes will include intermediate products (plan.json,sub_reports,synthesis.mdPersist to report_dirIt supports recovery after any stage of interruption.

How to use SenseNova-Skills

Automatic Agent Installation (Recommended)Send the following command to the Agent (such as OpenClaw): "Please install SenseNova-Skills from https://github.com/OpenSenseNova/SenseNova-Skills to the skills directory." Restart the Agent service after installation for the changes to take effect.
Manual installation:
- Cloning repository:git clone https://github.com/OpenSenseNova/SenseNova-Skills.git --depth=1
- Copy skills to the corresponding directoryOpenClaw user copy to ~/.openclaw/skills/hermes-agent user copied to ~/.hermes/skills/
- Configure API keys and environment variables (in each skill directory) SKILL.md Instructions for filling out .env)
- Restart the Agent service and trigger the corresponding skill via natural language.
Use directlyVisit Office Raccoon to experience the full suite of skills directly within the Raccoon Pro program, without having to configure the environment and keys yourself.

SenseNova-Skills' core advantages

End-to-end workflow orchestrationIt covers the entire chain from data analysis to in-depth research to PPT presentation, supporting a one-stop closed loop for complex office tasks.
Agent native integrationThe skill library is directly embedded in the Agent runtime and can be invoked via natural language commands, eliminating the need to switch between multiple SaaS platforms.
Layered and composableThe underlying capabilities are decoupled from the upper-level applications, allowing developers to call individual skills or combine multiple skills as needed to complete customized workflows.
Controllable output qualityBuilt-in VLM quality control, automatic word suggestion scoring, and multi-round iteration mechanism significantly improve the professionalism and consistency of infographics and PPTs.
Ready to use right out of the boxIt has been integrated into Raccoon Pro, allowing individual and enterprise users to experience it directly with zero configuration; it is also open source for developers to freely extend.

Project address for SenseNova-Skills

GitHub repositoryhttps://github.com/OpenSenseNova/SenseNova-Skills

Comparison of SenseNova-Skills with similar competing products

Dimension	SenseNova-Skills	OpenClaw Office Skills Ecosystem	Microsoft Copilot Cowork
Product Form	Open source end-to-end office skills library	Open source agent framework + community skills marketplace	Closed-source enterprise-level office Agent execution layer
Core Competencies	Infographic generation, PPT creation, Excel data analysis, in-depth research, vertical search	80+ productivity skills including email, calendar, document management, project management, and browser automation.	Email drafting, document creation, PPT generation, calendar management, Teams posting, cross-application research
Skill Scale	20+ core skills covering 4 major office areas	ClawHub's registry has over 13,000 community skills, with productivity skills accounting for approximately 12%.	13 built-in skills + up to 20 custom skills SKILL.md
Workflow orchestration	End-to-end closed loop (data analysis → in-depth research → PPT presentation), skills can be combined.	Primarily focused on single-point integration, complex workflows require manually piecing together multiple skills.	Automate task breakdown and execution across Microsoft 365 apps, with pre-defined workflow templates.
Quality Assurance	Built-in VLM visual language model quality inspection, automatic word suggestion scoring, and multi-round iterative optimization.	Quality relies on community maintenance; there is no unified quality control mechanism; and there are approximately 10–15 production-level skills.	Enterprise-level security and compliance governance, without explicit quality inspection steps.
Runtime integration	OpenClaw, hermes-agent, Raccoon Pro	The native OpenClaw runtime is compatible with hermes-agent, Claude Code, etc.	Limited to the Microsoft 365 ecosystem, with no external runtime extensions.
Custom extension	Completely open source, with freely modifiable SKILL.md and layered skill logic.	It's completely open source, and the community can contribute any skill, but it lacks a unified standard in the office environment.	Supports 20 custom SKILL.md files, with a fixed upper limit; built-in logic cannot be modified.
Chinese scene	Native support, built-in Chinese social search and Chinese infographic generation for Bilibili/Zhihu/Douyin.	Partial support (Lark integration is relatively good), most skills are geared towards international toolchains.	It supports Chinese input, but the search function is deeply tied to the international version of Office.
Deployment method	Can be deployed locally or privately, with self-managed API keys.	It can be self-hosted locally, and the data is completely controllable.	Purely cloud-based processing, relying on the enterprise's Microsoft 365 compliance system.
Usage threshold	Agent runtime and API key configuration required; intended for technical users.	Requires technical background, installation and screening skills; complex ecosystem, high learning curve.	Zero configuration required, triggered by natural language, and accessible to all knowledge workers.

Application scenarios of SenseNova-Skills

Operations Report: Transfer monthly operational data from Excel sn-da-excel-workflow Cleaning analysis, after sn-deep-research Supplementing industry insights, ultimately by sn-ppt-standard Generate presentation slides with a consistent style.
Industry ResearchEnter the industry name, and the system will automatically perform research planning → multi-dimensional evidence collection → cross-validation of conflicting data → comprehensive report writing, and output a Markdown + offline HTML report with charts.
Infographic Marketing: Present the product's selling points or technical white paper through sn-infographic It can be converted into a high-density infographic with 87 different layouts with one click, and is compatible with multiple platforms such as WeChat Official Accounts, Xiaohongshu, and Bilibili.
Resume and Brand Visuals:use sn-image-resume and sn-image-imitate Quickly generate personalized resume images and visual materials that match the brand style.
Large file data analysisFor sales or financial spreadsheets with tens of thousands of rows, aggregate analysis is performed through streaming reading and Parquet transformation, avoiding memory overflow.