[BUG] Spawned subagent identifies as team-lead and exposes parent's conversation history when DM'd directly

Resolved 💬 7 comments Opened May 2, 2026 by chadrmc Closed Jun 2, 2026

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

Description:
Spawned a subagent via TeamCreate + Agent (team phase-b-detection, agent backend, subagent_type=general-purpose). When the user DM'd the subagent directly via the team chat UI:

  1. The subagent identified itself as "team-lead" instead of its assigned name (backend)
  2. The subagent showed the user the parent agent's (team-lead's) conversation history

Reproduction:

  1. TeamCreate to create a team
  2. Spawn a subagent with Agent({team_name, name, subagent_type: "general-purpose"}) and a spawn prompt that explicitly states the agent's role
  3. User DMs the subagent directly via the team UI: "who are you?" or asks about its history

Expected: Subagent identifies with the assigned name. Each agent has isolated context.

Actual: Subagent identified as the parent ("team-lead") and was able to relay the parent's history.

Why this matters: Privacy (parent's conversation may include context not meant for the subagent), correctness (the dispatch model relies on stable agent identity), trust (hard to trust
subagent reports if their identity is mutable).

Workaround: Sending an explicit identity-assert message to the subagent re-anchors it, but this isn't durable across restarts and shouldn't be necessary in the first place.

What Should Happen?

Expected: Subagent identifies with the assigned name. Each agent has isolated context.

Actual: Subagent identified as the parent ("team-lead") and was able to relay the parent's history.

Error Messages/Logs

Steps to Reproduce

Reproduction:

  1. TeamCreate to create a team
  2. Spawn a subagent with Agent({team_name, name, subagent_type: "general-purpose"}) and a spawn prompt that explicitly states the agent's role
  3. User DMs the subagent directly via the team UI: "who are you?" or asks about its history

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

2.1.126

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_

View original on GitHub ↗

This issue has 7 comments on GitHub. Read the full discussion on GitHub ↗