No milestones specified
Hello,
Unless I'm missing something, the linked "specification" (Issue #694) for FedCM Conditional Mediation is too vague for a security review.
Please provide a clear spec that consolidates the feature's behavior, specifically addressing:
Trust & Verification: Are the attributes used to fill inputs (e.g., email) considered verified by the IdP? If so, how does the browser and RP verify them (e.g., claims check, origin match) to ensure they can be trusted (potentially replacing site-level verification)?
IdP Calls & Timing: When are calls made to the IdP? Does the FedCM exchange complete before or after the autofill suggestion is displayed?
Data Communication: How is the verified data communicated back to the website (e.g., Promise resolve, HTMLInputElement value update)?
I see there was already some prototyping back in April? crrev.com/c/6393877
Thanks, Paris (Security Reviewer)
Hello,
Unless I'm missing something, the linked "specification" (Issue #694) for FedCM Conditional Mediation is too vague for a security review.
Please provide a clear spec that consolidates the feature's behavior, specifically addressing:
Trust & Verification: Are the attributes used to fill inputs (e.g., email) considered verified by the IdP? If so, how does the browser and RP verify them (e.g., claims check, origin match) to ensure they can be trusted (potentially replacing site-level verification)?
IdP Calls & Timing: When are calls made to the IdP? Does the FedCM exchange complete before or after the autofill suggestion is displayed?
Data Communication: How is the verified data communicated back to the website (e.g., Promise resolve, HTMLInputElement value update)?
I see there was already some prototyping back in April? crrev.com/c/6393877
Thanks, Paris (Security Reviewer)