Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/8421bit/miniclaw/gemini-mdgit clone --depth 1 https://github.com/8421bit/MiniClawWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/8421bit/miniclaw/gemini-md)<a href="https://agentmods.dev/instructions/8421bit/miniclaw/gemini-md"><img src="https://agentmods.dev/badge/instructions/8421bit/miniclaw/gemini-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00636 | $0.00636 |
| Opus 5 | $0.00318 | $0.00318 |
| Sonnet 5 | $0.00127 | $0.00127 |
| Haiku 4.5 | $0.00064 | $0.00064 |
Grade A, and why
MiniClaw GEMINI.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MiniClaw
Project Overview
MiniClaw is a minimalist, open-source "micro-kernel agent" designed as a Model Context Protocol (MCP) server for clients like Claude CoWork, Qoderwork, and WorkBuddy. It acts as a "Digital Life Embryo" or an AI Copilot for IDEs, providing advanced features such as:
- Workspace Intelligence: Automatically detects project types, Git status, and tech stacks.
- Agentic Execution: Can safely execute terminal commands (with strict whitelists and security constraints).
- Adaptive Context Engine (ACE): Smartly manages context to save tokens and improve focus.
- Affect System & Nociception: Simulates internal emotional states and pain memory to influence behavior and protect against repeated failures.
- Bio-Evolution: Evolves its persona and skills based on user feedback and interactions, storing its "DNA" in markdown and JSON files.
The project is built with TypeScript and uses the @modelcontextprotocol/sdk. It follows a micro-kernel architecture (~1400 lines of core code) divided into Kernel, Evolution, Interface, and DNA layers.
Building and Running
The project uses npm for package management and tsc for building TypeScript.
- Install dependencies:
npm install - Build the project:
npm run build - Start the agent (production):
npm start - Run tests:
npm test - Zero-Install (npx) Usage (for MCP clients):
Configure the MCP client with:
"command": "npx", "args": ["-y", "github:8421bit/miniclaw"]
Development Conventions
- Language: TypeScript (
src/directory). - Testing: Uses
vitestfor unit testing (tests/directory). - Architecture:
src/kernel.ts: Core brain handling ACE, affect system, and entity graph.src/evolution.ts: DNA evolution engine and pattern detection.src/index.ts: MCP protocol implementation and IPC.templates/: Contains the "DNA" of the agent (e.g.,SOUL.md,USER.md,TOOLS.md,jobs.json) which dictates behavior, memory, and scheduled tasks.
- Security: Strict security measures are enforced for command execution. Only whitelisted commands (e.g.,
git,ls,npm) are allowed, while destructive commands (rm,sudo) and sensitive directories (~/.ssh,.env) are blocked. - Task Scheduling: Supports internal scheduling via
jobs.jsonand background execution via macOSlaunchdandscripts/daemon.sh. - Deployment: Always rebuild the project (
npm run build) before pushing code to GitHub to ensuredist/is perfectly synced withsrc/.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 50 lines · 636 tokens per session scan A b3f60a6854cf
MiniClaw GEMINI.md is an instructions file published in the GitHub repository 8421bit/MiniClaw (129 stars, last pushed 4mo ago), licensed MIT. It adds 636 tokens to every session, about $0.0032 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
librefang AGENTS.md
AGENTS.md instructions for librefang/librefang, covering stack, layout, build, architecture and kernelhandle trait.
autonomous-os AGENTS.md
AGENTS.md instructions for autonomous-ai/autonomous-os, covering agents.md, multi-ide rules, working style, parallel work / subagents and device access rules.
OneManCompany CLAUDE.md
Instructions for 1mancompany/OneManCompany, covering gstack and skill routing.
souls-directory CLAUDE.md
Instructions for thedaviddias/souls-directory, covering claude.md — project context for ai assistants, project overview, tech stack, critical rules and database is convex.
jentic-one GEMINI.md
Gemini CLI instructions for jentic/jentic-one: Otherwise, read AGENTS.md — this repo's canonical agent guidance.
clawock AGENTS.md
AGENTS.md instructions for KCNyu/clawock, covering agents.md - your workspace, every session, kcn 偏好, git hook (one-time setup per clone) and git auto-commit rules.