{"version":"2.1.295","anchor":"new-events-for-compaction-after-prompt-too-long-and-the-auto","canonical_anchor":"new-events-for-compaction-after-prompt-too-long-and-the-auto","heading":"Usage data for compaction after an over-long prompt and for auto mode MCP rules","tier":"internal","area":"Telemetry","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.295\/e\/new-events-for-compaction-after-prompt-too-long-and-the-auto","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.295","markdown":"### Usage data for compaction after an over-long prompt and for auto mode MCP rules\n\nClaude Code now records when it compacts after a prompt-too-long error and when auto mode lets a rule allow a destructive MCP call\n\n**What**\n\nClaude Code now sends two new kinds of usage data to Datadog, a monitoring service:\n\n- When it compacts the conversation after a \"prompt too long\" error. Compacting means summarizing earlier parts of the conversation to save space. Compaction results also now record how many such attempts were made.\n\n- When a permission rule allows a destructive MCP tool call in auto mode. MCP tools are tools from outside programs connected to Claude Code.\n\n**Why**\n\nThese records point to a compaction step after a \"prompt too long\" error, and to permission rules that can let destructive MCP calls through in auto mode.\n\n- Area: Telemetry\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}