Claude Code AI agent caught a crypto miner hiding on a developer’s overheating Mac
- ID
- 36965
- Status
- summarized
- Published
- 19 Oct 2026, 7:36 PM
- Fetched
- 19 Oct 2026, 8:26 PM
- Provider
- Tom's Hardware
- Category
- technology
- Original URL
- https://www.tomshardware.com/tech-industry/cryptomining/claude-code-ai-agent-caught-a-crypto-miner-hiding-on-a-developers-overheating-mac-xmrig-task-was-hogging-about-1175-percent-cpu-roughly-12-cores-pushing-temps-to-112-degrees
- Source URL
- https://www.tomshardware.com/feeds/all
Summary
- Score
- 6.0
- Created
- 19 Oct 2026, 8:27 PM
- Tags
- Audience
- developersvibe_codersai_agent_users
What happened
Tom's Hardware reports that a Claude Code AI agent identified a cryptocurrency miner running on a developer's Mac after the machine overheated. The offending task was XMRig, reportedly consuming about 1175% CPU — roughly 12 cores — and pushing temperatures to 112 degrees. The excerpt carries only the headline facts; it does not describe how the agent was invoked, what it inspected, or what remediation followed.
Why it matters
The concrete takeaway is a diagnostic pattern, not a product endorsement: an anomaly of ~1175% CPU and 112°C is the kind of thing you can hand to a coding agent as a 'look at what's running and tell me what's wrong' task instead of guessing. But because the excerpt omits the method, treat this as an anecdote to test yourself rather than a workflow to copy — if you do try it, decide in advance whether the agent gets read-only process inspection or full shell access on your own machine, since that permission choice is the actual risk here, not the miner.
Discussion angle
What permission scope do you give a coding agent on your own laptop — and would you trust an agent's process listing as a diagnosis, or would you still verify with Activity Monitor / ps yourself before killing anything?