Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
SenseNova-Skills provides a collection of specialized modules designed for automation builders creating AI agents. The library focuses on structured office tasks, enabling the transition from manual data handling to automated agentic workflows.
These skills allow agents to perform complex operations such as multi-step Excel data cleaning, handling large datasets via Parquet conversion for performance, and applying conditional formatting to analysis results.
WHO IT'S FOR
AI agent / automation builders
building AI-powered office assistants and workflows
Data analysts
automating multi-step Excel data cleaning and analysis
Business intelligence engineers
processing large Excel datasets with Parquet optimization
Operations specialists
comparing spreadsheets and marking anomalies with colors
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Follow the documented setup, then try a first task.
I’m using
Agent Pack for Windows
Installs: Installs the selected OpenClaw and/or hermes-agent runtime and the complete SenseNova-Skills bundle; skillPaths is a bounded selection from that bundle. · OpenClaw, hermes-agent
Before you start
Windows 10 22H2 or later, or Windows 11
WSL2 already installed
A SenseNova API key plus the matching regional Base URL and model name
Platform: Windows
Install WSL2 before running Agent Pack; installation is performed inside WSL2.
Download the latest Agent Pack file whose name ends in `-windows-x64.exe`, double-click it, and follow the wizard. Select OpenClaw, hermes-agent, or both as desired.
When prompted for the LLM provider, choose Custom (OpenAI-compatible). Enter the API key and the Base URL and model name from the same regional row: International uses `https://token.sensenova.ai/v1`; Mainland China uses `https://token.sensenova.cn/v1`; both list `sensenova-6.8-flash-lite`.
After setup, start the installed agent and ask: "List the available skills, with a one-line description for each." Confirm that it lists the SenseNova skills. The standalone sn-deepresearch-cli and sn-proactive-agent runtime require the additional setup described in their guides.
Installs: Installs the selected OpenClaw and/or hermes-agent runtime and the complete SenseNova-Skills bundle; skillPaths is a bounded selection from that bundle. · OpenClaw, hermes-agent
Before you start
Xcode Command Line Tools already installed
Homebrew already installed
A SenseNova API key plus the matching regional Base URL and model name
Platform: macOS
Complete the macOS platform prerequisites, Xcode Command Line Tools and Homebrew, before using Agent Pack.
Download the latest Agent Pack file whose name ends in `-macos-universal.pkg`, double-click it, and use the graphical wizard to select OpenClaw, hermes-agent, or both.
When prompted for the LLM provider, choose Custom (OpenAI-compatible). Enter the API key and the Base URL and model name from the same regional row: International uses `https://token.sensenova.ai/v1`; Mainland China uses `https://token.sensenova.cn/v1`; both list `sensenova-6.8-flash-lite`.
After setup, start the installed agent and ask: "List the available skills, with a one-line description for each." Confirm that it lists the SenseNova skills. The standalone sn-deepresearch-cli and sn-proactive-agent runtime require the additional setup described in their guides.
Installs: Installs the selected OpenClaw and/or hermes-agent runtime and the complete SenseNova-Skills bundle; skillPaths is a bounded selection from that bundle. · OpenClaw, hermes-agent
Before you start
A SenseNova API key plus the matching regional Base URL and model name
Platform: Linux
Download the latest Agent Pack file whose name ends in `-linux.sh`. In a terminal opened in the directory containing that downloaded file, run the documented command; `*` matches the release-specific portion of its filename.
When prompted for the LLM provider, choose Custom (OpenAI-compatible). Enter the API key and the Base URL and model name from the same regional row: International uses `https://token.sensenova.ai/v1`; Mainland China uses `https://token.sensenova.cn/v1`; both list `sensenova-6.8-flash-lite`.
After setup, start the installed agent and ask: "List the available skills, with a one-line description for each." Confirm that it lists the SenseNova skills. The standalone sn-deepresearch-cli and sn-proactive-agent runtime require the additional setup described in their guides.
Installs: Installs the selected OpenClaw and/or hermes-agent runtime and the complete SenseNova-Skills bundle; skillPaths is a bounded selection from that bundle. · OpenClaw, hermes-agent
Before you start
A SenseNova API key plus the matching regional Base URL and model name
Platform: Linux
Run the documented Agent Pack installer command in a Linux terminal.
When prompted for the LLM provider, choose Custom (OpenAI-compatible). Enter the API key and the Base URL and model name from the same regional row: International uses `https://token.sensenova.ai/v1`; Mainland China uses `https://token.sensenova.cn/v1`; both list `sensenova-6.8-flash-lite`.
After setup, start the installed agent and ask: "List the available skills, with a one-line description for each." Confirm that it lists the SenseNova skills. The standalone sn-deepresearch-cli and sn-proactive-agent runtime require the additional setup described in their guides.
Turn a Markdown report into a self-contained HTML page
Uses sn-md-to-html-report
Replace `[path/to/report.md]` with the Markdown report path.
Replace `[audience and purpose]` with who will read the page and what they should learn or decide.
Use sn-md-to-html-report to turn [path/to/report.md] into a self-contained HTML feature page for [audience and purpose]. Preserve the source's factual claims and conclusion strength, create plan.md before writing the HTML, and review the result for accessibility and responsive behavior.
Replace `[topic]` with the technology or problem of interest.
Repository search does not require a GitHub token; a token is mandatory only for code search.
Use sn-search-code to search GitHub for five repositories about [topic], using repository search rather than code search. Return the documented JSON results with each item's title, URL, and snippet.
Claude
OpenClaw
Hermes