Seahorse Agent is an engineering platform for enterprise knowledge, long-term memory, and governable agents. It is not a single chat demo, but a reference implementation for building enterprise-grade AI applications, integrating RAG, user memory, user profiles, Agents/Skills/Tools, tenant governance, auditing, and observability into one deployable, verifiable, and extensible system.
4 files for Claude Code and Codex: handoff, seahorse-consumer-web-exposure-audit, seahorse-memory-system-audit, seahorse-agent CLAUDE.md — 1,438 tokens loaded in every session.
CLAUDE.md A 1,438 tok .skills/handoff/SKILL.md A 19 tok .skills/seahorse-consumer-web-exposure-audit/SKILL.md A 76 tok .skills/seahorse-memory-system-audit/SKILL.md A 52 tok These files are onceMisery/seahorse-agent's own configuration — they tell Claude Code and Codex how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.