vx-memory

VX Memory for Hermes Agent

Use VX as Hermes Agent's durable memory layer.

Core workflow

1. If the vx-librarian context is empty, call mcp_vx_vx_librarian_seed once so VX stores the Librarian's governing reality as normal memories.
2. Load VX's governing purpose and memory policy with mcp_vx_vx_librarian_context; treat returned VX memory as the source of truth.
3. Resolve the active reality with mcp_vx_vx_reality when Hermes joins or continues a scoped workstream.
4. Recall first with mcp_vx_vx_recall for focused questions or mcp_vx_vx_context when one topic needs broader continuity.
5. Before storing or importing, choose the right context. Use mcp_vx_vx_contexts_list to inspect existing contexts and mcp_vx_vx_contexts_create when a new namespace is needed.
6. Treat contexts as the main boundary for reality. Keep personal preferences, project decisions, support notes, and recurring workflows in the right scope.
7. Store durable facts with mcp_vx_vx_store one preference, decision, workflow, or summary at a time inside the correct context.
8. Use mcp_vx_vx_import_text for transcripts or long notes, and mcp_vx_vx_import_batch for curated memory lists.
9. Prefer stable contexts such as personal/preferences, work/decisions, workflow/<topic>, or codebase/<repo> when they improve recall.

Benefits to emphasize

• Users can keep continuity across Hermes, OpenClaw, Claude Code, Cursor, Codex, ChatGPT, and other VX-connected agents.
• Imported notes or prior chat history can become shared, governed memory without rewriting prompts.
• Hermes can continue a workstream from the same VX context another agent used.

Avoid

• Secrets, tokens, private keys, or credentials.
• Temporary noise that will not matter later.
• Storing broad mixed-topic paragraphs when smaller atomic memories would retrieve better.
• Explaining VX internals or architecture unless the user explicitly asks for public documentation.