Making an LLM retype its own 64-hex auth token on every tool call works — until it improvises a tool and improvises the token too, and subagents spawned via sessions_spawn never got one at all. Fix: a native before_tool_call hook injects each agent's token host-side, keyed by ctx.agentId. Deterministic, per-agent scopes, the model never touches the secret.https://micelclaw.com/blog/agent-identity#selfhosted #ai #mcp #opensource
Making an LLM retype its own 64-hex auth token on every tool call works — until it improvises a tool and improvises the ...
