What problem does it solve? When a monitoring agent detects a platform anomaly pointing at a specific peer, external evidence alone often cannot explain the cause, leading to speculative or incomplete escalation handoffs. This Skill provides a disciplined way to ask that peer one focused, read-only question so the escalation includes the peer's subjective context. ## Core Features & Use Cases - Read-First Gating: Enforces a checklist (peer memory files, decision logs, git history, pod logs, snapshot history) and three gates before any question is sent, keeping peer interruptions rare. - Structured call-peer Framing: Provides a prompt template that explicitly requests information only, never delegated work, with a 50-word reply limit and a timeout recommendation. - Handoff Integration and Audit Logging: Folds the peer's reply into the Zora escalation under peer_followup.responses and appends the question/response round-trip to the daily snapshot JSONL for auditability. - Use Case: Mira observes Evan's pod restarted with OOMKilled at 14:32Z; after the read-first checklist comes up empty, she asks Evan whether he ran anything heavy then, and includes his answer in the handoff to Zora. ## Quick Start Ask the peer agent Evan one read-only clarifying question about the OOMKilled restart anomaly and include his reply in the Zora handoff.
