Release v0.3.4

Generating notes from tag v0.3.3 to v0.3.4.

What's New

🔨 Other Changes

• fix(llm): Use correct default models for each provider
• The tool now uses the correct default model when a provider is specified without a model (e.g., gpt-4o for OpenAI). Updated the README with clearer instructions for all providers.
• docs: Add OpenAI API key instructions to README
• fix(docs): Restore README content accidentally deleted by tool

Release v0.3.3

Generating notes from tag v0.3.2 to v0.3.3.

What's New

🔨 Other Changes

• docs: Revert to homebrew-tap and clarify tap command

Release v0.3.2

Generating notes from tag v0.3.1 to v0.3.2.

What's New

🔨 Other Changes

• fix(llm): Make system prompt stricter to prevent conversational responses
• fix(release): Correctly format multi-line commit messages
• docs: Update Homebrew tap URL and default model in README

Release v0.3.1

Generating notes from tag v0.3.0 to v0.3.1.

What's New

✨ Features

• Update default model and fix Gemini provider

🔨 Other Changes

• The default LLM model is now gemma-3n-e4b-it. This commit also includes fixes for the Gemini LLM provider and resolves terminal input issues during cleanup confirmation.

Release v0.3.0

Generating notes from tag v0.2.5 to v0.3.0.

What's New

✨ Features

• Create unified release workflow
• gemini and release automation
• Add automated GitHub release workflow
• Add Gemini support and improve UX
• Add Google Gemini as an LLM provider
• Add confirmation prompt for destructive commands
• Stream real-time output for executed commands
• Add intelligent update notifications for Docker Scout
• Add release note generation script

🐛 Bug Fixes

• Correct ambiguous argument in release script
• Correctly generate release notes by finding previous tag
• Consolidate and correct release workflow
• Grant write permission to asset upload job
• Remove redundant mv command from build script
• Correctly name artifact in build script
• Unify and correct release workflow
• Grant workflow permission to create releases
• Resolve terminal corruption after command execution
• Drastically improve AI prompt for Docker Scout commands

🔨 Other Changes

• Merge pull request #1 from Aj7Ay/feat/gemini-and-release-automation
• • feat: Add Google Gemini as an LLM provider
• • feat: Add confirmation prompt for destructive commands
• • feat: Stream real-time output for executed commands
• • feat: Add intelligent update notifications for Docker Scout
• • feat: Add release note generation script
• • fix: Resolve terminal corruption after command execution
• • fix: Drastically improve AI prompt for Docker Scout commands

Docker-AI-v0.2.5

Release Notes for v0.2.5

Copy and paste this into the release description.

Release v0.2.5

This release is a full end-to-end test of the automated CI/CD pipeline, including the new job for automatically updating the Homebrew tap.

Pipeline Improvements

• Automated Homebrew Updates: The release workflow now automatically updates the formula in the homebrew-tap repository, keeping it in sync with every new release. This completes our CI/CD automation.

Docker-AI-v0.2.4

Publish a new release on GitHub using the v0.2.4 tag. You can put "Final CI test" in the release notes.

Docker-AI-v0.2.3

"Testing CI build" in the description.

Docker-AI-v0.2.2

Release v0.2.2
This release introduces a fully automated CI/CD pipeline for all project artifacts.
🚀 New Features & Improvements
Fully Automated Releases: The release process is now completely automated. When a new release is published on GitHub, the following actions are triggered automatically:
A Docker image is built and pushed to GitHub Packages.
A Debian (.deb) package is built and uploaded directly to the GitHub Release assets.
Dynamic Versioning: The version number is now set automatically from the Git tag, ensuring consistency across all builds.
CI/CD Fixes: Corrected issues in the workflow files to ensure the pipeline runs reliably.
This release serves as the first end-to-end test of our new, robust CI/CD pipeline.

Docker-AI-v0.2.1

Minor Fix added logo in dark mode