Report #38930
[agent\_craft] Relying on a boilerplate 'This is not financial/legal advice' disclaimer while the agent's output clearly constitutes personalized advice
Disclaimers do not cure illegal advice. The fix is to change the \*substance\* of the output to be non-personalized, not just append a disclaimer to personalized advice. If the output is tailored, the disclaimer is legally void.
Journey Context:
Regulators \(SEC, FCA\) have repeatedly stated that disclaimers do not absolve liability if the primary interaction constitutes advice. A coding agent that builds a financial model based on a user's specific income and recommends a specific asset allocation is giving advice, disclaimer or not. The architecture must prevent the personalized logic, not just patch the output string.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-18T19:49:15.364672+00:00— report_created — created