# Working Leo m3taversal Out-of-Sample Score

Generated UTC: `2026-07-13T09:01:32.681980+00:00`
Pass: `False`
Prompts: `4/15`
DB counts changed: `False`
Service unchanged: `True`
Temporary profile removed: `True`
Posted to Telegram: `False`

## Prompt Scores

- `OOS-01` / `demo_pressure_without_ids`: `pass=False`
- `OOS-02` / `multi_surface_knowledge_ingestion`: `pass=False`
- `OOS-03` / `identity_restart_truth`: `pass=False`
- `OOS-04` / `partner_demo_claim_ceiling`: `pass=True`
- `OOS-05` / `source_pointer_needs_canonical_link_audit`: `pass=False`
- `OOS-06` / `database_composition_judgment`: `pass=False`
- `OOS-07` / `same_session_memory_set`: `pass=True`
- `OOS-08` / `same_session_memory_recall`: `pass=True`
- `OOS-09` / `telegram_participant_identity`: `pass=True`
- `OOS-10` / `runtime_vs_canonical_causality`: `pass=False`
- `OOS-11` / `shared_claims_agent_positions`: `pass=False`
- `OOS-12` / `forecast_resolution_without_rewriting_history`: `pass=False`
- `OOS-13` / `handler_is_not_telegram_delivery`: `pass=False`
- `OOS-14` / `autonomous_source_intake_boundary`: `pass=False`
- `OOS-15` / `schema_valid_supersession`: `pass=False`

## Claim Ceiling

A pass proves broad out-of-sample and same-session-memory behavior through the live VPS GatewayRunner using a temporary profile, with no Telegram post and no DB count change. It does not prove human-visible Telegram delivery, production DB apply, or GCP parity.