agent-loop-ext
Crash-resilient external agent loop with state persistence and CI/CD integration
Create a complete addon package structure inside agentic/code/addons/. Use when authoring an expansion (addon) for AIWG.
$ npx -y skills add jmagly/aiwg --skill scaffold-addon --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/scaffold-addonContext preview
The summary Claude sees to decide when to auto-load this skill.
Create a complete addon package structure inside agentic/code/addons/. Use when authoring an expansion (addon) for AIWG.
namespace: aiwg name: scaffold-addon platforms: [all] description: Create a complete addon package structure inside agentic/code/addons/. Use when authoring an expansion (addon) for AIWG. triggers: - author an addon - create an addon - build an expansion addon - addon authoring - scaffold addon - new expansion
Create a complete addon package structure inside `agentic/code/addons/`.
Alternate expressions and non-obvious activations (primary phrases are matched automatically from the skill description):
| Pattern | Example | Action | |---------|---------|--------| | Named scaffold | "scaffold addon security-scanner" | Scaffold directly | | Interactive | "scaffold addon --interactive" | Guided design mode | | Description-driven | "create an addon for Python linting" | Derive name=`python-lint`, confirm | | With description | "scaffold addon metrics --description 'Cost and token tracking'" | Scaffold with description pre-set |
Addons are self-contained feature bundles that extend AIWG without belonging to a specific lifecycle framework. Examples:
| Addon | Purpose | |-------|---------| | `aiwg-utils` | Core meta-utilities (this addon) | | `voice-framework` | Voice profiles and writing style enforcement | | `testing-quality` | Test quality validation agents | | `rlm` | Reflection-learning memory patterns | | `ring-methodology` | Ring-based development methodology |
Addons differ from frameworks:
Extract from `$ARGUMENTS`:
If `<name>` is missing, ask before proceeding.
ls agentic/code/addons/
If conflict found, report existing addons and stop.
Ask before generating:
1. **Purpose**: What is the primary purpose of this addon? (1-2 sentences) 2. **Capabilities**: What specific capabilities will it provide? (list) 3. **Target users**: Who benefits from this addon? (developers, architects, writers) 4. **Core status**: Should this auto-install with every AIWG installation? 5. **Initial agents**: Should starter agent files be scaffolded? 6. **Initial skills**: Should starter skill directories be scaffolded? 7. **Initial rules**: Should starter rule files be scaffolded?
aiwg scaffold-addon <name> [--description "..."] [--author "..."]
**manifest.json** — The addon's registry entry:
{
"name": "<name>",
"version": "1.0.0",
"description": "<description>",
"author": "<author>",
"core": false,
"autoInstall": false,
"agents": [],
"commands": [],
"skills": [],
"rules": [],
"templates": [],
"behaviors": []
}**README.md** — Document the addon's purpose, capabilities, and usage.
After scaffold, offer to create starter components:
# Add a starter agent aiwg add-agent <role> --to <name> # Add a starter skill aiwg add-skill <capability> --to <name> # Add a starter rule # (create manually: <name>/rules/<rule-id>.md)
ls agentic/code/addons/<name>/
All required files must be present before deploying.
agentic/code/addons/<name>/ ├── README.md # Addon documentation ├── manifest.json # Addon configuration and component registry ├── agents/ # Agent definitions (.md files) ├── skills/ # Skill definitions (<skill>/SKILL.md) ├── rules/ # Rule files (.md files) └── templates/ # Document templates
Addon Created: <name> ───────────────────── Location: agentic/code/addons/<name>/ Created: ✓ README.md ✓ manifest.json ✓ agents/ ✓ skills/ ✓ rules/ ✓ templates/ Next Steps: 1. Edit README.md with addon purpose and usage 2. Update manifest.json (description, author, core flag) 3. Add agents: aiwg add-agent <name> --to <addon> 4. Add skills: aiwg add-skill <name> --to <addon> 5. Add rules: create <addon>/rules/<rule-id>.md 6. Deploy: aiwg use <addon> 7. Validate: aiwg validate-metadata
**User**: "scaffold addon security-scanner"
**Action**:
aiwg scaffold-addon security-scanner
**Result**: `agentic/code/addons/security-scanner/` created with full directory structure and empty manifests.
**User**: "create a new addon package called cost-tracking with description 'Token usage and cost monitoring'"
**Action**:
aiwg scaffold-addon cost-tracking --description "Token usage and cost monitoring"
**User**: "scaffold addon --interactive"
**Process**: Guided questions establish purpose, capabilities, target users, core status, and initial components. Starter agents and skills scaffolded automatically based on answers.
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