What problem does it solve? AI agents lose track of who the user is between conversations, forcing users to repeat context every session. This Skill configures Honcho memory for Anakot so user preferences, facts, and patterns persist across sessions while each agent profile maintains its own isolated peer identity. ## Core Features & Use Cases - Cross-Session User Modeling: Honcho builds a user representation from observed messages and injects session summaries, representations, and peer cards into the system prompt automatically. - Multi-Profile Peer Isolation: Each Anakot profile gets its own AI peer sharing one workspace, so agents develop independent views while seeing the same user model. - Tunable Dialectic Engine: Three orthogonal knobs (cadence, depth, reasoning level) control how often, how many rounds, and how hard Honcho's dialectic reasoning runs, plus five bidirectional tools for profile, search, context, reasoning, and conclusions. - Use Case: You run separate "coder" and "writer" agent profiles. Use this Skill to give each its own Honcho peer via anakot profile create coder --clone, tune dialecticDepth to 2 for the coder, and cap contextTokens so long sessions never blow up the prompt. ## Quick Start Set up Honcho memory for my Anakot agent and show me the connection status and peer configuration.
