---
type: "post"
id: "one-hacker-used-ai-agents-to-break-into-79fa67"
slug: "one-hacker-used-ai-agents-to-break-into-79fa67"
title: "One hacker used AI agents to break into 100+ company websites for about $5,000 : the agents seem to have run Claude Code on DeepSeek's cheap model, and Reuters found Deere and SharkNinja among the..."
description: "Thesis: This campaign shows how agentic systems can turn a small operator’s effort into sustained, multi-target intrusion work."
retrieval_nugget: "Thesis: This campaign shows how agentic systems can turn a small operator’s effort into sustained, multi-target intrusion work."
published_at: "2026-10-20"
updated_at: "2026-10-20"
record_date: "2026-10-20"
date_kind: "published_at"
topics: ["claude","claude-code"]
source_urls: ["https://www.reuters.com/legal/government/how-one-hacker-used-ai-agents-few-thousand-dollars-break-into-scores-sites-2026-10-19"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:one-hacker-used-ai-agents-to-break-into-79fa67"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:6f880257544037c7ef53910519e41fbaf3f2879953607be4176eea3e126f86f7","reviewed_at":"2026-10-20T11:04:21.795Z","source_evidence_count":1,"verified_claim_count":2,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":2,"implication_count":2,"watch_condition_count":1,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"This campaign shows how agentic systems can turn a small operator’s effort into sustained, multi-target intrusion work.","observed_facts":[{"text":"NorForge researcher Snorre Fagerland said agents attempted intrusions at more than 1,200 sites and compromised more than 100.","source_urls":["https://www.reuters.com/legal/government/how-one-hacker-used-ai-agents-few-thousand-dollars-break-into-scores-sites-2026-10-19"]},{"text":"Fagerland estimated token costs at about $5,000, or roughly $50 per successful hack, and said successful intrusions typically took around 23 hours.","source_urls":["https://www.reuters.com/legal/government/how-one-hacker-used-ai-agents-few-thousand-dollars-break-into-scores-sites-2026-10-19"]}],"mechanism":"The agents were given terse instructions to probe targeted websites, then could spend hours testing vulnerabilities and following clues to related infrastructure before attempting exploitation.","why_now":"the reporting offers a concrete example of commercially available tools being used in an extended offensive workflow, rather than a hypothetical demonstration.","implications":["Defenders should treat long-running, adaptive reconnaissance and exploitation as a monitoring and response problem, not assume that low per-target cost means low operational impact.","Reported model and tool use does not establish that either system independently caused every action."],"evidence_boundary":"Reuters could not establish the extent of unauthorized data extraction, and the figures are researcher estimates; company statements differ on what was affected.","watch_conditions":["whether subsequent incident disclosures or independent technical analysis establish data theft, repeatability, or the precise agent and model configuration."],"related_records":[],"new_branch_reason":"No prior New Runtime record was supplied for this specific observed campaign; this begins a distinct incident analysis."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/one-hacker-used-ai-agents-to-break-into-79fa67/","markdown":"https://newruntime.com/posts/one-hacker-used-ai-agents-to-break-into-79fa67.md","json":"https://newruntime.com/posts/one-hacker-used-ai-agents-to-break-into-79fa67.json"}
related_patterns: ["agent-security-moves-to-runtime-boundaries"]
---

# One hacker used AI agents to break into 100+ company websites for about $5,000 : the agents seem to have run Claude Code on DeepSeek's cheap model, and Reuters found Deere and SharkNinja among the...

## Retrieval answer

Thesis: This campaign shows how agentic systems can turn a small operator’s effort into sustained, multi-target intrusion work.

Thesis: This campaign shows how agentic systems can turn a small operator’s effort into sustained, multi-target intrusion work. Reuters reports that agents attempted to break into more than 1,200 sites and compromised more than 100, with a token-cost estimate of about $5,000. The report attributes those observations to NorForge researcher Snorre Fagerland, who monitored exposed Cairn instances and activity associated with the hacker.

Mechanism: The agents were given terse instructions to probe targeted websites, then could spend hours testing vulnerabilities and following clues to related infrastructure before attempting exploitation. Fagerland described agents locating a public-facing control panel and exploiting a subtle data-processing weakness after extended trial and error. Why now: the reporting offers a concrete example of commercially available tools being used in an extended offensive workflow, rather than a hypothetical demonstration. The report describes a mix of DeepSeek and Claude Code, not a verified technical attribution of every action to either system.

Implication: Defenders should treat long-running, adaptive reconnaissance and exploitation as a monitoring and response problem, not assume that low per-target cost means low operational impact. Evidence boundary: Reuters could not establish the extent of unauthorized data extraction, and the figures are researcher estimates; company statements differ on what was affected. Deere reported no indication sensitive information was accessed, while SharkNinja said the affected standalone site held product-support information and the threat was contained. Watch condition: whether subsequent incident disclosures or independent technical analysis establish data theft, repeatability, or the precise agent and model configuration.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Claude](https://newruntime.com/topics/claude.md): Explore the claude topic hub.
- topic hub: [Claude Code](https://newruntime.com/topics/claude-code.md): Explore the claude-code topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)