| name | duplication-sweep |
| description | Gemini-native structural collision detection — scans all repos for duplicate definitions with 2M+ token grounding |
| capability | skill.duplication_sweep |
| node_name | node_skill_duplication_sweep_orchestrator |
| dot_color | 208 |
Duplication Sweep Skill (Gemini Edition)
Lightweight structural scan that leverages Gemini's massive context window to identify colliding definitions across the entire organization, including Drizzle tables, Kafka topics, and Python models.
Workflow
- Multi-Repo Scan: Gemini ingests code across all repositories to identify potential collision classes.
- Structural Checks:
- D1 (Database): Detects duplicate Drizzle table definitions across schema files.
- D2 (Events): Identifies conflicting producer claims for the same Kafka topic.
- D3 (Migrations): Detects duplicate migration prefixes.
- D4 (Models): Gemini performs a "whole-project" analysis to find cross-repo model name collisions outside of
omnibase_core.
- Verdict Generation: Categorizes findings into PASS, WARN, or FAIL based on structural risk.
- Integration: Findings are fed into the autopilot system for automated halt decisions.
Gemini Advantages
- Whole-Project Model Mapping: Unlike simple grep-based checks, Gemini can understand the relationship between models in different repositories and distinguish between intentional sharing and accidental collisions.
- Intelligent Deduplication: Better at identifying when a "duplicate" is actually a re-implementation or a legitimate cross-repo reference.
- Accurate Risk Assessment: More reliably determines if a collision occurs in a production codepath vs a test or fixture.
Arguments
--check D1,D2: Comma-separated check IDs to run.--json: Output machine-readable results.
