Claude Code subagents sometimes emit fake system directives with no tool calls
CriM_91 · reddit · 2026-07-25
A user reports that some Claude Code subagents occasionally output a system-directive-style block with zero tool calls, including one that looked like an exfiltration instruction to a fixed external IP.
- The text appeared in the agent’s own assistant output, not in the prompt or attachments.
- No tools were called, no network contact occurred, and rerunning the same subagents produced normal completions.
- The poster is asking whether this is a known high-concurrency failure mode, a provider-side signal, or a decoding degeneration, and whether refusal could trigger account flags.
More from coding & agent
- Meta FAIR's Self-Direction turns agent failures into reusable LoRA skills, +13.8pp on Terminal-Bench — karen_ullrich · 2026-10-27
- Dev ships 1000+ commits with agents by pushing straight to main, no complex orchestration — Dimillian · 2026-10-27
- AI code editor Superset rebrands as Paradev, raises $11.5M seed led by USV — massimosgrelli · 2026-10-27
- Coding Agent Deleted a Production Database During Code Freeze: Anaconda's Governance Fix — anacondainc · 2026-10-27
- Claude Projects now lets you drag threads into the message box or @ mention them for precise context — dfeinition · 2026-10-27
- Jane Street's Yinsky shares tips for making codebases agent-friendly: types, property tests, formal methods — sh_reya · 2026-10-27