recovery-assurance

What problem does it solve?

Only SQLite-backed runtime state and direct recovery verification count as durable assurance evidence. Workspace notes or one-off artifacts do not count as durable knowledge by themselves.

Use this skill for backup freshness, restoreability, and disaster-recovery confidence.

Do not use it for generic filesystem health or planned change work:

• use reliability_ops for disk pressure or failed jobs
• use change_lifecycle for backup configuration changes
• use incident_response when the system is already in an outage

This skill uses the shared SQLite kernel via skill_key=recovery_assurance.

Core Features & Use Cases

• Tracks backup coverage and restore evidence using a structured graph model (backup_coverage, restore_evidence, rpo_gap, dr_runbook).
• Leverages helper scripts to collect, bootstrap, store, and query recovery evidence into a durable SQLite kernel.
• Provides guardrails and operator feedback contracts to manage recovery risk and articulation of gaps.

Quick Start

Execute recovery_assurance to initiate a conservative backup-coverage and restore-evidence assessment using the bundled helper scripts.