Skip to content
Back to skills

Smith Agenticdef

ASecurity

Generate or update the agentic environment definition for AgentSmith, SkillSmith, and CommandSmith

  • 207 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added May 27, 2026
data-aigobashgitdocumentation

Works with

  • claude code
  • cursor

Security analysis

A100/100

Scanned May 27, 2026

npx -y skills add jmagly/aiwg --skill smith-agenticdef --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Smith Agenticdef?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Smith Agenticdef
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/jmagly-smith-agenticdef/badge)](https://www.skillsdirectory.com/skills/jmagly-smith-agenticdef)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
namespace: aiwg
name: smith-agenticdef
platforms: [all]
description: Generate or update the agentic environment definition for AgentSmith, SkillSmith, and CommandSmith
commandHint:
  argumentHint: '[--verify-only] [--update] [--interactive] [--guidance "text"]'
  allowedTools: 'Read, Write, Glob, Bash'
  model: haiku
  category: smithing
---

# Agentic Definition Generator

Generate the agentic environment definition file that describes the platform's capabilities for creating agents, skills, and commands.

## Purpose

This command probes the current agentic platform and generates `.aiwg/smiths/agentic-definition.yaml` which is used by:
- **AgentSmith** - To know available models, tools, and agent deployment paths
- **SkillSmith** - To know skill structure and deployment paths
- **CommandSmith** - To know command categories and deployment paths

## Arguments

| Argument | Description |
|----------|-------------|
| `--verify-only` | Check existing definition without updating |
| `--update` | Update existing definition with any changes |

## Workflow

### 1. Detect Platform

Determine the agentic platform by checking for platform-specific directories:

```
Platform Detection Priority:
1. .claude/         → Claude Code
2. .factory/        → Factory AI
3. .github/agents/  → GitHub Copilot
4. .cursor/         → Cursor
5. .codex/          → OpenAI Codex
```

### 2. Probe Platform Capabilities

**For Claude Code:**
- Check `.claude/agents/` exists
- Check `.claude/commands/` exists
- Check `.claude/skills/` exists
- Verify CLAUDE.md or project context

### 3. Identify Available Models

Check platform configuration for supported models:
- `haiku` - Fast, simple tasks
- `sonnet` - Balanced (default)
- `opus` - Complex reasoning

### 4. Catalog Available Tools

Standard tool set:
- File: Read, Write, MultiEdit
- Search: Glob, Grep
- Execution: Bash
- Web: WebFetch, WebSearch
- Orchestration: Task, TodoWrite

### 5. Generate Definition

Create `.aiwg/smiths/agentic-definition.yaml`:

```yaml
# Agentic Environment Definition
# Generated by /smith-agenticdef
# Last updated: {timestamp}

platform:
  provider: claude
  version: "{detected_version}"
  detected_at: "{timestamp}"

agent_config:
  models:
    - haiku
    - sonnet
    - opus
  default_model: sonnet
  tools:
    available:
      - Read
      - Write
      - MultiEdit
      - Bash
      - Glob
      - Grep
      - WebFetch
      - WebSearch
      - Task
      - TodoWrite
    orchestration_default:
      - Task
      - Read
      - Write
      - Glob
      - TodoWrite

skill_config:
  supported: true
  structure: directory
  filename: SKILL.md

command_config:
  supported: true
  categories:
    - sdlc-management
    - sdlc-orchestration
    - development
    - utilities
    - smithing

deployment_paths:
  agents: .claude/agents/
  skills: .claude/skills/
  commands: .claude/commands/

capabilities:
  can_create_agents: true
  can_create_skills: true
  can_create_commands: true
```

### 6. Create Directory Structure

Ensure Smith directories exist:

```bash
mkdir -p .aiwg/smiths/agentsmith/specs
mkdir -p .aiwg/smiths/skillsmith/specs
mkdir -p .aiwg/smiths/commandsmith/specs
```

### 7. Initialize Empty Catalogs

Create catalog files if they don't exist:
- `.aiwg/smiths/agentsmith/catalog.yaml`
- `.aiwg/smiths/skillsmith/catalog.yaml`
- `.aiwg/smiths/commandsmith/catalog.yaml`

## Output

```
Agentic Definition Generated
============================

Platform: Claude Code
Version: 2025.12

Agent Configuration:
  Models: haiku, sonnet, opus
  Tools: 10 available
  Deployment: .claude/agents/

Skill Configuration:
  Supported: Yes
  Structure: directory/SKILL.md
  Deployment: .claude/skills/

Command Configuration:
  Supported: Yes
  Categories: 5
  Deployment: .claude/commands/

Files Created:
  ✓ .aiwg/smiths/agentic-definition.yaml
  ✓ .aiwg/smiths/agentsmith/catalog.yaml
  ✓ .aiwg/smiths/skillsmith/catalog.yaml
  ✓ .aiwg/smiths/commandsmith/catalog.yaml
```

## Verify Mode

With `--verify-only`, check existing definition:

```
Verifying Agentic Definition
============================

Definition: .aiwg/smiths/agentic-definition.yaml
Status: Valid

Platform Match: ✓ claude
Agent Paths: ✓ .claude/agents/ exists
Skill Paths: ✓ .claude/skills/ exists
Command Paths: ✓ .claude/commands/ exists

Catalogs:
  AgentSmith: ✓ 0 artifacts
  SkillSmith: ✓ 0 artifacts
  CommandSmith: ✓ 0 artifacts
```

## Error Handling

| Error | Resolution |
|-------|------------|
| No platform detected | Create `.claude/` directory or specify platform |
| Missing deployment directory | Create directory or check permissions |
| Invalid existing definition | Use `--update` to regenerate |

## References

- @$AIWG_ROOT/docs/smithing/agentic-smiths.md - Agentic Smiths documentation
- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/agents/agentsmith.md - AgentSmith agent
- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/agents/skillsmith.md - SkillSmith agent
- @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/agents/commandsmith.md - CommandSmith agent

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.