# AI use cases for Minimal Local Agent (MuzeAnisichael/minimal-local-agent)

Repo: [MuzeAnisichael/minimal-local-agent](https://github.com/MuzeAnisichael/minimal-local-agent) | Use cases: 3 | Page: https://dedupedai.com/use-cases/muzeanisichael-minimal-local-agent/

Minimal Local Agent is a small Python agent kernel with bounded tools, reviewable edits and a local SQLite audit log, usable from a CLI, a local web console or your own Python host with Ollama or a compatible API.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Run a local agent on one sandboxed folder: https://dedupedai.com/use-cases/muzeanisichael-minimal-local-agent/run-local-agent-on-sandboxed-folder/
- Embed a read-only agent in your own Python app: https://dedupedai.com/use-cases/muzeanisichael-minimal-local-agent/embed-read-only-agent-in-python/
- Learn how an agent loop works from a small codebase: https://dedupedai.com/use-cases/muzeanisichael-minimal-local-agent/study-minimal-agent-loop/

## How to run a local agent on one sandboxed folder with Minimal Local Agent

Repo: [MuzeAnisichael/minimal-local-agent](https://github.com/MuzeAnisichael/minimal-local-agent) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Point a local Ollama model at a single workspace with bounded tools and edits you preview before they write.

**Who it is for:** A developer who wants agent help on files without giving a model free rein on the disk.

**The problem:** General agents can touch anything and are hard to audit afterwards.

**The outcome:** An agent confined to one workspace, with every run recorded in SQLite.

### Steps

1. Clone with `git clone https://github.com/MuzeAnisichael/minimal-local-agent.git` and create a Python 3.11 venv
2. Install with `python -m pip install -e .` and pull a model with `ollama pull qwen3.5:9b`
3. Copy the config with `cp agent.example.toml agent.toml` and run `minimal-agent doctor`
4. Open the console with `minimal-agent web` and give it a task

### Prompt to paste into your agent

```text
Set up MuzeAnisichael/minimal-local-agent with Ollama, run `minimal-agent doctor`, then ask it to summarize the files in <workspace>. Confirm the run appears in the local SQLite audit history and no file changed.
```

### Caveat

Web tasks can only read or preview; approving an actual write requires the interactive CLI.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/muzeanisichael-minimal-local-agent/run-local-agent-on-sandboxed-folder/

## How to embed a read-only agent in your own Python app with Minimal Local Agent

Repo: [MuzeAnisichael/minimal-local-agent](https://github.com/MuzeAnisichael/minimal-local-agent) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README

Import AgentRuntime, set write_policy to deny, and run tasks that return a response plus a receipt hash.

**Who it is for:** A backend developer adding an agent step to an internal tool who needs a tamper-evident record.

**The problem:** Most agent frameworks are too large to embed and audit.

**The outcome:** A few lines of Python that run the agent and return a verifiable receipt.

### Steps

1. Install the package into your project
2. Load settings with Settings.load("agent.toml") and replace write_policy with deny
3. Call runtime.run with your task
4. Store outcome.response and outcome.receipt_hash

### Prompt to paste into your agent

```text
Using minimal_local_agent's AgentRuntime with write_policy="deny", run "List the workspace files" against <workspace> and print the response and receipt_hash. Verify the run made no writes.
```

### Caveat

Memory, planning and delegation exist only in the unreleased v1.1.0; the published release is v1.0.0.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/muzeanisichael-minimal-local-agent/embed-read-only-agent-in-python/

## How to learn how an agent loop works from a small codebase with Minimal Local Agent

Repo: [MuzeAnisichael/minimal-local-agent](https://github.com/MuzeAnisichael/minimal-local-agent) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README

Read and run one model and tool loop with policy and budgets to understand agents end to end.

**Who it is for:** A developer new to agents who finds big frameworks too opaque to learn from.

**The problem:** Large agent frameworks hide the core loop under layers of abstraction.

**The outcome:** A working mental model of runtime, tools, policy and audit state.

### Steps

1. Clone the repo and install it in a venv
2. Run the demo with `python examples/web_demo.py`
3. Read the runtime, tools and policy modules
4. Change a budget in agent.toml and rerun to see the effect

### Prompt to paste into your agent

```text
Clone MuzeAnisichael/minimal-local-agent, run `python examples/web_demo.py`, and explain the loop from host to runtime to tools to SQLite state. Verify your explanation against the source files you cite.
```

### Caveat

The web UI is currently in Chinese.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/muzeanisichael-minimal-local-agent/study-minimal-agent-loop/