Releases: 0xJord4n/aixion
v1.2.1-hotfix-typing-error
🚀 aixion v1.2.1: Type Definition Hotfix
🐛 Bug Fixes
- Fixed type definition error in provider configuration
- Resolved type validation issues affecting messages
📝 Details
This hotfix addresses a typing error that was causing validation issues in messages. The fix ensures proper type checking and validation.
🔍 Technical Changes
📦 Files Modified
src/input.ts
🎯 Impact
This update ensures:
- ✅ Correct type validation
💡 Notes
- No breaking changes introduced
- Backwards compatible with existing configurations
- All functionality remains unchanged
📚 Documentation
For more information, visit our GitHub repository
v1.2.0-file-input-support
🚀 aixion v1.2.0: File Input Support
✨ New Features
- Added file-based input support with
prompt_file
andsystem_file
- Added
content_file
support in messages array
📝 Details
This release introduces comprehensive file input support, allowing users to read prompts, system messages, and message content directly from files. This enables better organization of prompts and easier management of large-scale AI interactions.
🔍 Technical Changes
{
"provider": "openai",
"provider_options": {
"api_key": "your-api-key",
"timeout": 30000, // <- New timeout option
"max_retries": 3 // <- New retry option
},
"prompt_file": ".github/prompts/analysis.txt", // <- New file input
"system_file": ".github/prompts/system.txt", // <- New file input
"messages": [
{
"role": "user",
"content_file": ".github/prompts/query.txt" // <- New file input in messages
}
],
"model": "gpt-4",
}
📦 Files Modified
README.md
src/action.ts
src/input.ts
src/provider.ts
src/types.ts
🎯 Impact
This update enables:
- ✅ File-based input management
- ✅ Mixed content support (direct + file-based)
💡 Notes
- No breaking changes introduced
- Backwards compatible with existing configurations
- File paths are relative to repository root
- UTF-8 encoding used for all file operations
- Files must exist and be readable
- Can mix direct content and file-based content in messages array
📚 Documentation
For more information, visit our GitHub repository
v1.1.0-file-output-support
🚀 aixion v1.1.0: File Output Support
✨ New Features
- Added
save_path
configuration option for saving AI responses to files - Implemented automatic directory creation for output paths
- Added file output support across all AI providers
📝 Details
This release introduces the ability to save AI responses directly to files, enabling better integration with existing workflows and automated processing pipelines.
🔍 Technical Changes
{
"provider": "openai",
"provider_options": {
"api_key": "your-api-key"
},
"prompt": "Your prompt here",
"model": "gpt-4",
"save_path": "output/response.txt" // <- New configuration option
}
📦 Files Modified
README.md
src/index.ts
src/input.ts
src/action.ts
🎯 Impact
This update enables:
- ✅ Direct file output of AI responses
- ✅ Automatic directory creation
- ✅ Persistent storage of responses
- ✅ Enhanced workflow automation
💡 Notes
- No breaking changes introduced
- Backwards compatible with existing configurations
- UTF-8 encoding used for all file operations
- File paths are relative to workspace directory
- Existing files will be overwritten
📚 Documentation
For more information, visit our GitHub repository
v1.0.2-change-action-runtime
🚀 aixion v1.0.2-change-action-runtime
🔧 Runtime Changes
- Updated Node.js runtime compatibility
📝 Details
This version addresses runtime compatibility issues
🔍 Technical Changes
📦 Files Modified
package.json
(added new dependency)
🎯 Impact
This update ensures:
- ✅ Better compatibility with modern Node.js features
- ✅ More reliable action execution
- ✅ Consistent behavior across different GitHub Actions runners
💡 Notes
- No breaking changes introduced
- Enhanced cross-platform compatibility
📚 Documentation
For more information, visit our GitHub repository
v1.0.1-docker-workdir-fix
🚀 aixion v1.0.1
🔧 Bug Fixes
- Fixed incorrect working directory configuration in Docker setup
📝 Details
The primary fix addresses a critical issue in the Docker configuration where the working directory was improperly set, potentially causing build and execution failures.
🔍 Technical Changes
WORKDIR /app # <- Fixed working directory configuration
📦 Files Modified
Dockerfile
🎯 Impact
This update ensures:
- ✅ Correct file path resolution
- ✅ Proper build process execution
- ✅ Reliable container operations
💡 Notes
- No breaking changes introduced
- Backwards compatible with existing workflows
- Improves overall stability of containerized execution
📚 Documentation
For more information, visit our GitHub repository
v1.0.0
aiction v1.0.0
Initial release of aiction - A GitHub Action for AI integration.
Features
- Multi-provider support (OpenAI, Anthropic, Groq, Mistral, etc.)
- Three prompting methods:
- Simple prompt
- System + prompt
- Messages array (chat format)
- Configurable model parameters including temperature, max tokens, and stop sequences
- Custom API endpoints and headers
- Comprehensive output handling (text, usage stats, finish reason)
Usage
See the README.md for detailed installation and configuration instructions.
Supported Providers
- OpenAI
- Anthropic (Claude)
- Groq
- Mistral
- Cohere
- DeepInfra
- Fireworks
- Together AI
- XAI