Generate structured memory-sharpening workflows for implicit conversation changes.
npx skills add https://github.com/ejentum/integrations --skill ejentum-memory
Please help me install this Agent Skill. Skill: ejentum-memory Source: https://github.com/ejentum/integrations/tree/main/claude-code/skills/ejentum-memory Command: npx skills add https://github.com/ejentum/integrations --skill ejentum-memory
This Skill helps you reliably notice when a conversation has shifted in meaning, intent, or emotional state across turns—especially when changes are implicit rather than explicitly corrected.
Call this memory Skill for your next turn by describing the three most important changes you noticed (content change, delivery change, and what’s absent) and asking it to sharpen your perception.
High-intent search queries and answers about installing and using this skill.
To detect subtle tone shifts and emotional incongruence in multi-turn conversations, you can use a structured memory-sharpening workflow that applies a perception-first protocol to identify implicit changes. This method requires providing a raw observational query to generate a structured observation procedure.
Implicit state tracking is a conversation intelligence mechanism that identifies when meaning, intent, or emotional state shifts across turns without explicit correction. You need it when what changed matters more than what was said, such as detecting omissions or stale-state conditions.
To update your internal model when a user's plan implicitly pivots, you generate a state-updating observation procedure by describing the content change, delivery change, and what is absent. This structured approach reduces projection and helps you adjust what you believe is currently true.
The best way to avoid clinging to stale context during multi-turn dialogue is to apply a two-pass perception-first workflow that forces you to observe first and then sharpen your perception. This reduces projection and improves verification of implicit plan pivots or tone shifts.
No specific dependencies or components are required to use this multi-turn perception approach. You only need to provide a raw observational query describing the changes you noticed and use the memory mode call while following the perception-first protocol.
You should not use a memory-sharpening workflow for state tracking when changes in the conversation are explicitly corrected by the user, as this tool is specifically designed to detect implicit changes like subtle delivery shifts, omissions, and emotional incongruence where what changed matters more than what was said.
jfoxxAnalyze web page sections to identify content sequence breaking points.
hoavdcMap policies and evidence against a compliance framework to generate a prioritized remediation plan.
jose-compuCapture dialogue learnings into structured markdown logs for improvement.
DojoGenesisWrite structured memory entries to tiered markdown files with timestamps and metadata.
