跳至主要内容
小龙虾小龙虾AI
🤖

Stdio Skill

Stdin/stdout file inbox/outbox bridge for passing files to/from Clawdbot using an MCP stdio server. Use when you want a simple filesystem-backed dropbox: accept files into an inbox, move to tmp for processing, and emit deliverables to an outbox (or a specified path).

下载1.2k
星标0
版本1.0.0
AI 智能体
安全通过

技能说明


name: stdio-skill description: "Stdin/stdout file inbox/outbox bridge for passing files to/from Clawdbot using an MCP stdio server. Use when you want a simple filesystem-backed dropbox: accept files into an inbox, move to tmp for processing, and emit deliverables to an outbox (or a specified path)."

stdio-skill

Implement and use a local MCP stdio server that provides a simple inbox/outbox workflow backed by directories on disk.

Paths (workspace-relative):

  • stdio/inbox/ — user drops inputs here
  • stdio/tmp/ — scratch area (move/copy inputs here for processing)
  • stdio/outbox/ — put deliverables here for pickup

Start the MCP server (via mcporter)

This repo config should include an MCP server named stdio-skill.

  • List tools:
    • mcporter list stdio-skill --schema --timeout 120000 --json

Tooling model

Prefer:

  1. stdio-skill.stdio_list to see what’s waiting.
  2. stdio-skill.stdio_read (base64) to pull file contents.
  3. stdio-skill.stdio_move to move an item to tmp once you’ve claimed it.
  4. Write outputs with stdio-skill.stdio_write (base64) into outbox unless the user provided an explicit destination path.

No deprecated aliases: use the stdio_* tools only.

Notes

  • This skill is intentionally dumb/simple: it does not interpret file formats.
  • It is safe-by-default: operations are restricted to the three directories above.
  • For large files: prefer passing by path + moving files, not embedding giant base64 blobs in chat.

如何使用「Stdio Skill」?

  1. 打开小龙虾AI(Web 或 iOS App)
  2. 点击上方「立即使用」按钮,或在对话框中输入任务描述
  3. 小龙虾AI 会自动匹配并调用「Stdio Skill技能完成任务
  4. 结果即时呈现,支持继续对话优化

相关技能