Is Smoke Monkey an AI model?

No. Smoke Monkey is not an LLM. It is an embeddable, framework-agnostic runtime that turns LLMs (NVIDIA, Gemini, OpenAI, Claude, Ollama) into autonomous engineering agents capable of multi-step planning, calling tools, editing files, and running test loops with zero runtime dependencies.

Do I need a database (Postgres, SQLite, or Redis)?

No database is required. State is persisted through lean session files on disk, or you can plug in your own custom storage adapter. You can deploy it as a CLI, a desktop app, or inside a serverless function without provisioning a database.

How does human-in-the-loop permission gating work?

Smoke Monkey comes with three permission policies: allow-all, deny-all, and ask-default (or a custom function resolver). When a sensitive tool like edit_file or run_command is called under "ask", the runtime emits a permission.required event, pausing execution until your code or UI calls agent.resolvePermission(id, "allow" | "deny").

How does the Model Context Protocol (MCP) server work?

Smoke Monkey ships an official stdio MCP server in smoke-monkey-harness-mcp. Point Claude Code, Cursor, Codex, or opencode at it using npx -y smoke-monkey-harness-mcp to let other AI tools plan, scaffold, verify, and run agents using Smoke Monkey automatically.

What are SKILL.md skills and how do they prevent context bloat?

Skills are directories containing a SKILL.md workflow file. The system prompt only includes a one-line catalog of available skills. When a task requires specialized expertise (e.g. frontend-ui-engineering, test-driven-development, or security-hardening), the agent calls use_skill to load instructions just-in-time, saving thousands of tokens.

What is an AI Agent Harness?

An AI agent harness is the execution environment that bridges raw Large Language Models (LLMs) to real-world software engineering capabilities. While an LLM generates tokens, an AI harness manages deterministic phase loops (explore, plan, edit, verify, recover), tool execution sandboxing, token budget compaction, and safety gates needed for autonomous problem-solving.

What makes Smoke Monkey an Agentic AI runtime rather than a simple wrapper?

Unlike basic prompt-response chains, Smoke Monkey implements true agentic AI architecture: a deterministic 6-phase state machine with automatic failure recovery, runaway loop guards, 24 built-in developer tools, dynamic subcontext isolation, and interactive human-in-the-loop permission pauses.

Is Smoke Monkey Harness 100% Open Source AI?

Yes. Smoke Monkey Harness is 100% open source under the permissive MIT license with zero runtime dependencies. There is no telemetry lock-in, no mandatory cloud backend, and no database requirement—giving engineering teams complete sovereignty over their agent code.

What Agent APIs does Smoke Monkey Harness provide?

Smoke Monkey provides composable TypeScript Agent APIs: createAgent() for initialization, agent.run() for multi-turn execution, agent.on() for fine-grained lifecycle and streaming events, agent.resolvePermission() for interactive human authorization, subcontext switching APIs, and stdio Model Context Protocol (MCP) server endpoints.

Can I run completely offline without API keys?

Yes! Smoke Monkey natively supports Ollama out of the box. Set provider: "ollama" and model: "qwen3:8b" with workspacePath, and the agent will run locally on your machine with zero external network requests.

Safety & Control

6 min readUpdated: October 2026

Human-in-the-Loop Permission Gating in AI Agents: Safe Autonomous Execution

Unconstrained AI agents with terminal access can execute dangerous commands (like `rm -rf` or unintentional git force pushes). Smoke Monkey Harness introduces a robust, interactive permission gating system with three explicit policies: allow-all, deny-all, and ask-default.

Technical Review: Smoke Monkey Core Architecture Team

|Tested on Node.js 18+ & Bun|TypeScript 5.x

Quick Answer & Executive Definition

Human-in-the-Loop Permission Gating in AI Agents: Safe Autonomous Execution: Unconstrained AI agents with terminal access can execute dangerous commands (like `rm -rf` or unintentional git force pushes). Smoke Monkey Harness introduces a robust, interactive permission gating system with three explicit policies: allow-all, deny-all, and ask-default. Designed as a zero-dependency, open-source TypeScript architecture under the MIT License with native Model Context Protocol (MCP) support and deterministic phase state machines.

Key Architectural Takeaways

The 3 Policies: Configure granular permissions per tool (allow, ask, deny).
Interactive Pause Protocol: Emits `permission.required`, suspending the loop until approved or rejected.
Web & CLI Ready: Easily bind approval dialogs to React components or terminal prompts.
Safe Defaults: Read-only tools auto-allow while file mutations and shell commands ask for confirmation.

The Dangers of Ungated Tool Execution

When building autonomous agents that interact with real systems, allowing unrestricted shell execution is a major security hazard. Prompt injection or hallucinated commands can lead to data loss. Smoke Monkey Harness treats tool execution as a gated capability.

How the Pause Protocol Operates

When a tool with the "ask" policy is called, Smoke Monkey creates a pending permission request and pauses the execution loop. The host application receives a permission.required event containing the tool name, arguments, and a unique call ID.

permission-listener.tstypescript

1agent.on('permission.required', async (event) => {
2  const { toolCallId, toolName, args } = event.data;
3  console.log(`⚠️ Agent wants to run: ${toolName} with args:`, args);
4
5  // Present to user in UI or terminal
6  const userApproved = await askUserConfirmation();
7  await agent.resolvePermission(toolCallId, userApproved ? 'allow' : 'deny');
8});

Google Search Questions & Answers

Frequently Asked Questions

Q:What happens if a permission request is denied?

The harness feeds a structured cancellation response back into the LLM context, allowing the agent to transition to its `recover` phase and formulate an alternative approach without crashing.

Build with Smoke Monkey Harness

Zero dependencies. 24 built-in tools. Human-in-the-loop safety. 100% open source under the MIT License.

npm install smoke-monkey-harness

Human-in-the-Loop Permission Gating in AI Agents: Safe Autonomous Execution

The Dangers of Ungated Tool Execution

How the Pause Protocol Operates

Frequently Asked Questions

Q:What happens if a permission request is denied?

Related Alternatives & Comparisons

ChatGPT Agent APIs Alternative: Open Source TypeScript Harness

LangChain TypeScript Alternative: Zero Dependencies & Deterministic Loops

Build with Smoke Monkey Harness