Banks scramble over unauthorized AI agents in accounts; devs push MCP-based approval rails
yenkel · x · 2026-10-22
Per Axios, Morgan Stanley, Bank of America, JPMorgan and Wells Fargo are racing to contain AI agents accessing customer accounts with user-granted credentials — allowed by users but violating banks' human-login rules. Executives call it horrifying, fearing agents with purchase or fund-transfer authority. yenkel argues banks should build a paved road via WebMCP/MCP with per-action approval customization instead of blocking agents entirely, or they'll lose business.
More from coding & agent
- Free book 'LangGraph in Small Pieces' teaches agents from blocks to production — truckkicker · 2026-10-22
- Cloudflare's AI security-audit skill tops GitHub with 38.5k stars — alex_verem · 2026-10-22
- Roboclaw agent autonomously traces and fixes OpenClaw UI bug, PR merged — heyneighbor · 2026-10-22
- Bruin open-sources LLM inference as a first-class primitive in data pipelines — karakanb · 2026-10-22
- YC's Design Review: Magic Patterns demos MagicOS canvas and dissects why startup sites feel 'vibe coded' — ycombinator · 2026-10-22
- Skills over MCP plus progressive disclosure: fixing context bloat in domain-specific AI systems — sebgehr · 2026-10-22