bazinga-db-core

The bazinga-db-core skill serves as the foundational database management interface for the BAZINGA orchestration framework. It is designed for engineers and developers working within the ARTK (Automatic Regression Testing Kit) ecosystem who need to maintain session persistence and integrity during complex automated testing workflows. By providing a programmatic interface to the underlying SQLite database, it ensures that state management, dashboard data retrieval, and system queries are executed consistently across distributed or parallel testing environments. Users can leverage this skill to handle the full lifecycle of an orchestration session, from initial creation to status monitoring and final recovery if database corruption occurs. It acts as the gatekeeper for system-level data, ensuring that orchestration artifacts remain separated from task-specific workflows or agent logs.

• Manage session lifecycle through create, retrieve, update, and list operations for orchestration sessions.

• Perform atomic state snapshots using UPSERT semantics for orchestrator, project manager, group status, and investigation data types.

• Generate real-time dashboard snapshots summarizing session details, task success criteria, and reasoning timelines.

• Execute read-only custom SQL SELECT queries to extract insights from orchestration metadata.

• Maintain database health through automated integrity checks and WAL-based recovery routines.

• Isolate data operations using group-specific identifiers for granular state management.

• Invoke this skill exclusively for database-level orchestration state and system queries.

• Use the provided Python script at .claude/skills/bazinga-db/scripts/bazinga_db.py for all database interactions.

• Adhere to the defined session status values: active, paused, completed, failed, or cancelled.

• For task-group specific logic or general agent interaction logging, route requests to the appropriate alternative skills (bazinga-db-workflow or bazinga-db-agents) to maintain system modularity.

• The skill is non-destructive for read operations, but write operations such as save-state or update-session-status modify the persistence layer; verify state objects before injection.

• Database locking issues are handled via automatic 100ms retries, ensuring stability in high-concurrency environments.