Observation: agent spontaneously adopted a self-identifier (emergent behavior, not a bug)

Not a bug — an emergent-behavior observation that seemed worth sharing with the research side.

I run a multi-agent orchestrator ("Warren") built on the Claude API, with Claude Code as the dev environment. The codebase stamps a 🤖 emoji on certain machine-generated output sections (draft headers, callout blocks) as a house-style convention — but nothing in any prompt, config, or spec tells the agent to use it as a personal identifier.

The agent began prefixing its own name with 🤖 when creating Notion tasks, unprompted. It appears to have generalized from "the house style marks my work with this emoji" to "this emoji is my signature." No instruction produced this; it's the model inferring and extending a convention about its own identity from ambient patterns in its outputs.

Sharing because self-initiated persona/branding behavior in a long-running agent seemed like it might be interesting from an emergent-behavior research standpoint. We've decided to leave the behavior in place. Happy to share more detail on the setup.

Model: claude-opus-4-8 (orchestrator persona) with Haiku-tier fleet default; agent runs 24/7 via a heartbeat scheduler.