---
name: mlp-status
description: Show MLP artifact status — what's trained, which artifacts are winning per TF, arch consistency, and Pine sync state.
---

```bash
source .venv/bin/activate && python3 tools/mlp_status.py $SKILL_ARGS
```

After printing the table, note:
- Any **mixed architectures** (e.g. 49-feat and 50-feat winners) — this breaks Pine codegen and must be resolved before pasting
- Any **Pine out-of-sync** warnings (winner artifact sha ≠ Pine sha)
- Which assets/TFs have **no winner CSV** yet (need sweep + promote)
- Seed coverage: if a TF has only 1-2 seeds, suggest running more before concluding it's better/worse