agent-loop-ext
Crash-resilient external agent loop with state persistence and CI/CD integration
Create a new AIWG addon with AI-guided setup
$ npx -y skills add jmagly/aiwg --skill devkit-create-addon --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/devkit-create-addonContext preview
The summary Claude sees to decide when to auto-load this skill.
Create a new AIWG addon with AI-guided setup
namespace: aiwg name: devkit-create-addon platforms: [all] description: Create a new AIWG addon with AI-guided setup
Create a new standalone addon with AI assistance.
/devkit-create-addon <name> [--interactive]
| Argument | Required | Description | |----------|----------|-------------| | name | Yes | Addon name (kebab-case recommended) |
| Option | Description | |--------|-------------| | --interactive | Enable interactive mode with guided questions |
agentic/code/addons/<name>/ ├── manifest.json # Addon configuration ├── README.md # Documentation ├── agents/ # Agent definitions ├── commands/ # Slash commands └── skills/ # Auto-triggered skills
When `--interactive` is specified, I will ask:
1. **Purpose**: What is the primary purpose of this addon? 2. **Capabilities**: What specific capabilities should it provide? 3. **Initial agents**: Should I create any starter agents? 4. **Initial commands**: Should I create any starter commands? 5. **Core status**: Should this be a core addon (auto-installed)?
# Quick creation /devkit-create-addon security-scanner # Interactive guided creation /devkit-create-addon code-metrics --interactive
1. **Validate name**: Ensure name follows kebab-case convention 2. **Check existence**: Verify addon doesn't already exist 3. **Gather info**: In interactive mode, ask clarifying questions 4. **Generate manifest**: Create manifest.json with appropriate fields 5. **Create structure**: Build directory structure 6. **Generate README**: Create documentation with usage instructions 7. **Optionally create components**: If interactive, offer to create initial agents/commands 8. **Report success**: Show created files and next steps
For non-interactive creation, you can also use:
aiwg scaffold-addon <name> --description "..." --author "..."
Reusable project context and specialist workflows for the AI tools you already use. Plan software, coordinate specialist reviews, prepare campaigns, investigate incidents, organize research, curate media, and maintain operational knowledge.
Repo: jmagly/aiwg
Crash-resilient external agent loop with state persistence and CI/CD integration
Detect requests for iterative autonomous agent loops and route to the appropriate loop executor
Automatically execute tests when code-generating agents modify source files, enforcing the execute-before-return pattern
Enable agent loops to learn from similar past tasks and share patterns across loops
Query and manage the executable feedback debug memory
Execute tests on generated code and iterate until passing