nova-mind

Automated pipeline that processes chat messages to extract memories, facts, and vocabulary

Communicate with peer agents and spawn subagents in the NOVA agent ecosystem. Use when you need to delegate work to another agent, message a peer (Graybeard, Newhart), spawn a subagent (Coder, Scribe, Gidget, etc.), or determine how to reach a specific agent. Covers agent identification, peer messaging via agent_chat, subagent spawning with correct model/thinking config, and domain-based routing.

Review recent activity and extract institutional knowledge back into system artifacts. Captures missing workflow steps, implementation details into SKILL.md files, tool notes into TOOLS.md, and process lessons into memory. Trigger on: "introspect", "review what we did", "capture learnings", "update workflows from recent work", "what did we miss", "refine our processes", or after completing a significant multi-step task.

Communicate with peer agents and spawn subagents in the NOVA agent ecosystem. Use when you need to delegate work to another agent, message a peer (Graybeard, Newhart), spawn a subagent (Coder, Scribe, Gidget, etc.), or determine how to reach a specific agent. Covers agent identification, peer messaging via agent_chat, subagent spawning with correct model/thinking config, and domain-based routing.

Implement semantic search over agent memory using vector embeddings. Use when building AI memory systems, enabling meaning-based recall, or setting up proactive context retrieval. Covers PostgreSQL pgvector, OpenAI embeddings, and memory search patterns.

Manage a private Certificate Authority for mTLS authentication. Use when creating CA infrastructure, signing client certificates, configuring nginx for mutual TLS, or managing certificate lifecycle. Covers OpenSSL CA operations, CSR signing, and certificate verification.