Dear User

Dear User scans the configuration of an AI assistant and generates a written report that mimics a conversation with the agent. It examines the user’s setup to identify repeated patterns, leaking keys, and drifting components, then provides a letter‑style summary that includes three scores for collaboration, system health, and security. The tool also offers concrete suggestions, such as code snippets, to address the issues it finds.

The security aspect looks for exposed API keys across a range of services, detects contradictory rules, and highlights orphaned jobs or overlapping skills that could cause silent failures. It pulls warnings from platforms like GitHub, Supabase, npm, and Vercel, and flags secret types found in instructions, memory, hooks, and skills.

Additional features include an onboarding conversation to create initial CLAUDE.md or AGENTS.md files, a visual “Spotify Wrapped” style overview of agent‑human interaction, and a history view that tracks score trends over time. The tool is positioned for developers who integrate AI agents into their workflows and want automated insight into collaboration quality, health, and security.