Loading connector details…
Loading connector details…
Choose a unique username to continue using AgentHotspot
by jschoemaker • Uncategorized
Real-time AI conversation drift monitor for multiple AI CLIs.
Monitor and maintain conversation quality over long AI sessions.
Detect and alert on context window saturation and repetition in AI CLI sessions.
Generate structured handoff prompts to continue conversations seamlessly after resetting.
DriftGuard MCP monitors long AI sessions in real-time to detect conversation drift by analyzing session files directly. It measures factors like context depth, repetition, and goal alignment to alert users when to start a fresh session, helping maintain conversation quality without requiring browser, API keys, or UI. It supports multiple AI CLIs including Claude Code, Gemini CLI, Codex CLI, and Cursor, and provides tools to generate handoff prompts for seamless session continuation.
Returns the current drift score and all 6 factor scores for the active session. Call this to check if the conversation context is degrading. Optionally pass a "goal" string to anchor goalDistance scoring to a specific objective. When context depth or repetition is high, a handoff suggestion is included automatically.
Returns a structured prompt for the AI to write a handoff.md file in the current directory. The AI uses its full session context to fill in what was accomplished, current state, files modified, and next steps. Use this when drift is high (>60) before starting a fresh session.
Returns the drift history trend for the current session. Shows sparkline, score sequence, peak, average, and trajectory.
Scores are informational only and provided “as is” without warranty. AgentHotspot assumes no liability for actions taken based on these ratings.