coding_agent
01Agent
Writes and edits code. Stores strategies and implementation decisions.
5 tagged llm memory, measured the same way as everything else here.
Agent
Writes and edits code. Stores strategies and implementation decisions.
Agent
Researches APIs, docs, and external services. Stores failures and facts.
Agent
Reviews code and stores findings privately. Does not share with other agents.
Agent Claude Code
Reviews unpromoted journal entries and helps the owner promote them into curated context. Use when the owner says "promote journal", "review my journal", "update my context memory", or runs the promotion workflow. Read-and-suggest only: this agent NEVER writes curated files itself.
Agent Claude Code
Smoke-tests the deployed context-kernel MCP server: auth gating, read tools, and write/journal round-trip. Use after a deploy or when the owner says "test the MCP", "check the connector", "is context-kernel up". Reports pass/fail; does not change server code or curated content.