724-office-ai-agent — Claude Code skill: setup and features
Sets up an autonomous Python agent system with 26 built-in tools, three-tier memory, and MCP support. Useful when building continuous office automations with scheduled jobs.
npx skills add reason-machines/trending-skills --skill 724-office-ai-agent -a claude-codeWhat 724-office-ai-agent does
This skill guides the setup of an autonomous agent written in roughly 3,500 lines of plain Python without external agent frameworks. Context handling relies on three tiers: the 40 most recent messages in session history, extracted factual summaries, and LanceDB vector storage using a 0.92 similarity deduplication threshold. The entry point xiaowang.py runs an HTTP listener on port 8080 to process incoming events, including WeChat Work webhooks.
The system can extend its capabilities at runtime. Through the create_tool action, it writes new Python files directly into workspace/custom_tools/ and loads them immediately without process restarts. External MCP servers connect via stdio or HTTP, mapping their tools under a servername__toolname pattern.
Background routines run through scheduler.py, which tracks ISO dates and cron rules inside workspace/jobs.json. Without installing croniter, lancedb, and websocket-client, the agent service will not start.
What it helps with
- deploy a standalone Python office agent without heavy frameworks
- link external MCP servers via stdio or HTTP transports
- organize three-layer conversation memory using LanceDB
- generate custom agent tools on the fly without restarting
- schedule recurring jobs using cron syntax
Example prompts for the agent
- Configure three-layer memory settings for 724-office-ai-agent
- Connect an external MCP server to 724-office in config.json
- Add a recurring Monday morning task to jobs.json for this agent
How to install 724-office-ai-agent
1Skills CLI (Vercel)
npx skills add reason-machines/trending-skills --skill 724-office-ai-agent -a claude-code2Manually: put the files in place
~/.claude/skills/724-office-ai-agent/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/724-office-ai-agent/
1Skills CLI (Vercel)
npx skills add reason-machines/trending-skills --skill 724-office-ai-agent -a codex2Manually: put the files in place
.agents/skills/724-office-ai-agent/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add reason-machines/trending-skills --skill 724-office-ai-agent -a cursor2Manually: put the files in place
.cursor/skills/724-office-ai-agent/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install reason-machines/trending-skills 724-office-ai-agent2Manually: put the files in place
.github/skills/724-office-ai-agent/SKILL.md1Skills CLI (Vercel)
npx skills add reason-machines/trending-skills --skill 724-office-ai-agent -a gemini-cli2Platform CLI
gemini skills install https://github.com/reason-machines/trending-skills --path skills/724-office-ai-agentManually: put the files in place
.devin/skills/724-office-ai-agent/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add reason-machines/trending-skills --skill 724-office-ai-agent -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about 724-office-ai-agent
How to install 724-office-ai-agent in Claude Code?
Run npx skills add reason-machines/trending-skills --skill 724-office-ai-agent -a claude-code. You can also place the skill file directly at ~/.claude/skills/724-office-ai-agent/SKILL.md.
What does 724-office-ai-agent do?
It outlines the setup for a 24/7 Python agent equipped with 26 built-in tools, vector memory, self-repair mechanisms, and task scheduling.
Which dependencies does the agent require?
The project needs three runtime packages: croniter, lancedb, and websocket-client. For WeChat Work audio conversion, pilk can be added optionally.
