jin-boCommunity·9 Agent Skills Included

agentao

Governed local runtime with permissions, memory, and plugins

Runs language-model workflows locally with strict permission modes, tool confirmation, and sandboxing. Eliminates uncontrolled command execution, lost context, and hidden reasoning through memory, compaction, and full session replay. Connects to MCP servers, sub-agents, and plugins so hosts can embed one consistent runtime in CLIs or Python apps.

94|7

↗

npx skills add jin-bo/agentao --all -g -y

Available:

Tells the coding agent how to build, test, and navigate this repository's architecture, including package management with uv, lint gates, and the permission and tool system design.

All Skills in This Repository (9)

Pure Emerald Level Indicators

📦 In Repo

jin-bo

matplotlib-charts

Generate standardized PNG charts with matplotlib for headless workflows.

duckdb-analyst

Query local Parquet files with DuckDB and return SQL statements.

daily-digest

Compile daily markdown digests from GitHub commits and Hacker News stories.

ocr

Extract text from image files using the Qwen VL OCR model.

pro-ppt

Generate standardized 16:9 2K business presentation slides from speech drafts.

zootopia-ppt

Convert presentation scripts into 3D animated anthropomorphic animal slides.

kanban-tasks

Break down feature requirements into atomic task cards with acceptance criteria.

support-triage

Automate inbound support ticket triage and routing with policy enforcement.

skill-creator

Automate creation, testing, and optimization of Claude Code skills.

Community

Advanced

Frequently Asked Questions

FAQPage Schema

How to install Agentao?▼

Run `npx skills add jin-bo/agentao --all -g -y` in your terminal to install all skills in this suite globally.

What does Agentao do?▼

It is a local-first Python runtime that runs model-driven workflows with permission modes, tool confirmation, persistent memory, and session replay. You can use it as an interactive CLI or embed it in your own Python app.

How does Agentao keep automated runs safe?▼

Every tool call passes through a permission engine with read-only, workspace-write, full-access, and plan modes, plus optional macOS sandboxing. Nothing executes without matching your configured rules.

Can Agentao connect to external services?▼

Yes. It supports MCP servers over stdio, HTTP, and SSE, plus an ACP protocol for full agent-to-agent communication and a plugin system for extensions.

Which model providers work with Agentao?▼

Any OpenAI-compatible provider works, including Anthropic, Gemini, and DeepSeek. Set the API key, base URL, and model name in your environment or .env file.