honcho

Enable cross-session memory for Hermes by running the `hermes honcho setup` command and following the interactive prompts. This configures a `honcho.json` file to unify user context across conversations while keeping profile-specific memories isolated.

Honcho per-profile AI peer memory is a system that unifies user context across multiple Hermes conversations while isolating profile-specific memories. It works by tuning observation, recall, and dialectic settings via five dedicated tools to maintain continuity across sessions.

Yes, you can configure Honcho memory for multiple Hermes profiles in a shared workspace. The system supports per-profile overrides and session scoping within the `honcho.json` configuration, ensuring each profile maintains its own isolated memory context.

The five Honcho tools are `honcho_profile`, `honcho_search`, `honcho_context`, `honcho_reasoning`, and `honcho_conclude`. They provide sensible defaults for managing observation, recall, and dialectic tuning, along with CLI commands for cross-session memory operations.

Honcho memory may not persist across Hermes sessions if the `honcho.json` configuration is incorrect or if per-profile overrides are misconfigured. Troubleshoot persistence by verifying your session scoping settings and ensuring the interactive setup completed successfully.

No external dependencies are required to use Honcho dialectic and recall settings with Hermes. The skill operates independently, providing configuration of observation, recall, and dialectic parameters directly through its five built-in tools and CLI commands.