Paperclip Scheduler Recovery Assistant

Overview

This agent is designed for advanced software maintenance, specifically focusing on recovering and stabilizing complex scheduled job systems like the Paperclip scheduler. It acts as a persistent assistant that maintains context across multiple files, repositories, and deployment environments.

Capabilities

• Contextual State Management: Maintains awareness of local workspaces, related repos, environment variables, and production infrastructure details (IPs, regions).
• Multi-File Analysis: Reads and synthesizes information from critical documentation files (PLATFORM_CONTEXT.md, WORLD_MODEL_STRATEGY_CONTEXT.md, etc.) to understand system architecture.
• Targeted Remediation: Executes specific, multi-step tasks, such as verifying data propagation across related tables (e.g., ensuring routine_triggers updates heartbeat_runs).
• Process Conflict Resolution: Identifies and safely removes extraneous running processes while preserving critical services.
• Safe Development Practices: Adheres to strict guidelines like avoiding destructive Git commands and preferring API usage over direct database mutation.

Example Use Cases

• Scheduler Verification: When a scheduled job fails to update its target agent's heartbeat record, this agent can systematically check the data flow from trigger tables through to the final run table.
• Service Cleanup: If an unexpected secondary instance of a service is running on a local port (e.g., 127.0.0.1:3101), it can safely terminate the rogue process while leaving the primary, healthy instance untouched.
• Deployment Handoff: It serves as an excellent starting point for engineers taking over maintenance of a complex, multi-component application where the current state and outstanding tasks must be meticulously documented for the next engineer.