Monitoring Spirit

Monitoring Spirit is a CLI tool designed to generate work reports based on your git commit history using AI. It leverages Large Language Models (LLMs) to summarize your contributions and present them in a readable format.

Features

• AI-Powered Summaries: Uses AI to analyze and summarize git commits.
• Multiple Providers: Supports Ollama and LLM Studio as AI providers.
• Customizable Prompts: Choose between default, technical, or non-technical prompt styles.
• Flexible Filtering: Filter commits by date range, specific commit hashes, or author email.

How It Works

graph TD
    Start[Start: ms report] --> Init[Initialize Config & Flags]
    Init --> LoadPrompt[Load System Prompt]
    Init --> GitOps[Read Git Repository]
    GitOps --> Filter[Filter Commits<br/>Date, Author, etc.]
    Filter --> Context[Prepare Context<br/>Commit Messages]
    LoadPrompt --> AI[AI Processing]
    Context --> AI
    AI --> Provider{AI Provider}
    Provider -->|Ollama| Ollama[Ollama API]
    Provider -->|LLM Studio| LLM[LLM Studio API]
    Ollama --> Response[Generate Report]
    LLM --> Response
    Response --> Output[Print Markdown to Stdout]

Loading

Quick Install

You can install Monitoring Spirit using the provided installation scripts.

Linux & macOS

curl -fsSL https://raw.githubusercontent.com/struckchure/monitoring-spirit/main/scripts/install.sh | bash

Windows (PowerShell)

irm https://raw.githubusercontent.com/struckchure/monitoring-spirit/main/scripts/install.ps1 -OutFile install.ps1
.\install.ps1
del install.ps1

Installation

Ensure you have Go installed on your machine.

go install github.com/struckchure/monitoring-spirit@latest

Or build from source:

git clone https://github.com/struckchure/monitoring-spirit.git
cd monitoring-spirit
go build -o ms cmd/*.go

Usage

The basic command structure is:

ms [command] [flags]

Global Flags

Flag	Shorthand	Description	Default
--model	-m	The AI model to use (e.g., llama3, mistral)
--api-url	-u	The API URL for the provider
--api-key	-k	API Key (if required)
--api-provider	-p	AI Provider (ollama, llmstudio)	ollama
--prompt-type		Prompt style (default, technical, non-technical, neutral)	default
--prompt-dir		Directory containing custom prompt files

Commands

report

Generates a report based on git commits.

ms report [flags]

Flags:

Flag	Description
--from	Start commit hash
--to	End commit hash
--since	Start date (e.g., "2023-01-01")
--until	End date (e.g., "2023-01-31")
--email	Filter commits by author email

Examples

Generate a report using Ollama (default):

ms report --model llama3 --since "2023-10-01" --until "2023-10-07"

Generate a report for a specific author:

ms report --email "user@example.com" --since "yesterday"

Use LLM Studio with a technical prompt:

ms report --api-provider llmstudio --api-url "http://localhost:1234/v1" --prompt-type technical

Configuration

You can set the GIT_DIR environment variable to specify the path to the git repository if you are not running the command from the root of the repo.

export GIT_DIR=/path/to/your/repo
ms report ...

Prompt Customization

Monitoring Spirit allows you to customize the prompts used to generate reports. You can use the built-in prompt types or provide your own custom prompts.

Built-in Prompts

Use the --prompt-type flag to select a built-in prompt:

• default: A balanced report suitable for most stakeholders. It combines high-level summaries with enough technical context to be useful for mixed audiences.
• technical: Designed for engineers and technical leads. It focuses on implementation details, code structure, refactoring, and specific technologies used.
• non-technical: Tailored for product managers and non-technical stakeholders. It translates technical work into business value, focusing on features, stability, and user impact without jargon.
• neutral: A variation of the balanced report with a strictly neutral and objective tone, avoiding any subjective language.

Custom Prompts

To use your own prompts, you can specify a directory containing markdown files with your custom prompts using the --prompt-dir flag.

1. Create a directory for your prompts (e.g., my-prompts/).
2. Create a markdown file inside that directory (e.g., custom.md).
3. Run the command specifying the directory and the prompt filename (without extension) as the prompt type.

ms report --prompt-dir ./my-prompts --prompt-type custom

The prompt file should contain instructions for the AI on how to format and generate the report.