[ctx]ctxseccontext security: attacks on the harness and context layer of AI agents, plus exploited-now news, hunting and playbooks.

← AI · markdown

official (Wells Fargo upcoming Online Access Agreement, Section 15, effective Nov 24, 2026); other banks' positions per Hunton Andrews Kurth legal summary · updated

Wells Fargo's new online terms, effective Nov 24, deem anything an AI agent does with your login to be authorized by you, 'at your own risk'; Chase's Jan 1 terms say much the same

  • Security angle: an agent holding your bank password looks exactly like you to the bank, bypasses MFA by design, and makes prompt-injected or compromised-agent payments hard to dispute as fraud.
  • See the tip on keeping bank credentials out of AI agents.

Sources: Wells Fargo: Online Access Agreement, upcoming version effective Nov 24, 2026, Section 15 (Third-Party Artificial Intelligence Agents or Agentic AI) · Hunton Andrews Kurth: Major Banks Move to Block or Limit Customer-Directed AI Agents (Nov 3, 2026) · Citi Online User Agreement, updated November 2026, Section H: Use of Third-Party AI Agents