Skip to content
Get packs

Self-Customize Agent

Lets an AI agent modify its own setup: install packages, add MCP servers, or edit its code and instructions.

Source Download skill
Repository
nanocoai/nanoclaw
Stars
31,072
Licence
MIT
Files
1 file

What it does

This skill teaches an AI agent that runs inside the NanoClaw container how to change its own environment when asked. Simple changes, like saving notes or standing instructions, can be made directly. System packages and new MCP servers need an admin's approval first, and changes to the agent's source code are handed to a separate 'builder' agent so a human can review the result before it is relied on.

Great for

  • Developers running a NanoClaw agent who want it to extend its own tools over time
  • Operators who want changes to the agent's code or packages to go through a review step
  • Teams experimenting with self-modifying agents in a sandboxed container

Skip it if

  • Anyone wanting a plain, fixed agent: the skill lets the agent change its own code and installs, which adds moving parts and risk
  • Environments without an admin approval step for installs or MCP servers, since the safety gate relies on that approval

Try asking

  • “Can you add a tool that reads RSS feeds to yourself?”
  • “Can you install ffmpeg so you can transcribe voice messages?”
  • “Add a standing instruction to always reply in British English.”
  • “Change how you format your daily summaries in the source code.”
Step-by-step guide: how to use Self-Customize AgentInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    self-customize.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip self-customize.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip self-customize.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Keep approval required for install_packages and add_mcp_server, and review each request before approving it.
  • Ask the agent to prototype in its workspace first, so one-off experiments don't change the container image.
  • Review the builder agent's change summary before accepting source edits, especially for files outside the requested scope.

What's inside

  • SKILL.md

Redistributed unmodified from nanocoai/nanoclaw under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.