| name | omg-predict |
| description | 5 expert personas debate proposed changes before implementation. Catches architectural, security, performance, and UX issues early. Use before major features or risky refactors. |
Codex Port Notice
This skill was ported from upstream reference material. Interpret command names, paths, and agent-routing guidance as Codex/Oh My Game Kit equivalents. Prefer active Codex tools and local project instructions over Codex-specific mechanics when they conflict.
Oh My Game Kit Predict — Expert Debate
5 expert personas review a proposed change and debate its merits, risks, and alternatives.
Usage
omg-predict "Add caching layer to API responses"
Personas
| # | Persona | Focus | Looks For |
|---|
| 1 | Architect | System design, scalability, coupling | Over-engineering, tight coupling, wrong abstractions, missing extensibility |
| 2 | Security | Attack surface, data flow, trust boundaries | Auth bypass, injection, data exposure, insufficient logging |
| 3 | Performance | Bottlenecks, memory, latency | N+1 queries, unnecessary allocations, cache invalidation issues |
| 4 | UX/DX | User/developer impact, edge cases | Breaking changes, poor error messages, confusing APIs |
| 5 | Devil's Advocate | Assumptions, alternatives, hidden costs | "What if this is the wrong approach entirely?" |
Process
- Read the proposed change description
- Each persona reviews independently:
- State their concerns (1-3 bullet points)
- Rate risk for their domain (1-5 scale)
- Suggest mitigations
- Synthesize:
- Consensus risk score (average of 5 ratings)
- Prioritized concern list (highest risk first)
- Go/No-Go recommendation with conditions
Output Format
## Prediction: {change description}
### Architect (Risk: 3/5)
- Concern: ...
- Mitigation: ...
### Security (Risk: 2/5)
- Concern: ...
- Mitigation: ...
### Performance (Risk: 4/5)
- Concern: ...
- Mitigation: ...
### UX/DX (Risk: 1/5)
- Concern: ...
- Mitigation: ...
### Devil's Advocate (Risk: 3/5)
- Alternative: ...
- Hidden cost: ...
### Consensus
- **Overall Risk: 2.6/5**
- **Recommendation: GO with conditions**
- **Top 3 concerns to address:**
1. ...
2. ...
3. ...
Anti-Rationalization Guards
| Trap | Reality |
|---|
| "This change is too simple for a debate" | Simple changes with hidden complexity cause the worst incidents. |
| "We already discussed this" | Discussion ≠ structured risk assessment. |
| "The user wants to start coding" | 5 minutes of prediction saves 5 hours of debugging. |
When to Use
- Before any change affecting >3 files
- Before any security-sensitive change
- Before any architectural decision
- When the team disagrees on approach
- When risk feels "medium" or higher
Sub-Agent Fork Hygiene
Sub-agent forking: see .agents/skills/omg-architecture/references/fork-hygiene.md.
