Friday, September 04, 2009

Disclosure standards and why they're critical

If you've read this blog over the last couple of years you've likely made note of the varying degrees of success I've had disclosing vulnerabilities.
You've seen the best of breed in AppRiver and SmarterTools.
You've also seen lessons in how to not handle disclosures in the likes of American Express and Ameriprise. I believe Ameriprise is Pwnie-worthy for Lamest Vendor Response given that Benjamin Pratt, Ameriprise’s vice president of public communications, said "There's no one at risk here." and that there are no plans to review any of the mechanisms the company may have in place to receive notifications from the public about website vulnerabilities. Wow. The Consumerist clarified those statements aptly with "we assume he means, "No one important on our side of things. Our customers can suck it."

I take disclosure very seriously. I believe it is a deep seated, inherent responsibility that rests squarely on the shoulders of vendors and site operators. Equally, disclosure must be responsible, even when efforts to advise the vendor have come up empty. To that end:, the result of a recent interview I gave to by Ira Victor of the Data Security Podcast on the topic of mishandled disclosures. We decided on a joint project, thus

Report Security Flaws exists to increase awareness and responsiveness in Internet vendors and web site operators when they receive security-related disclosures.
It is our hope that all vendors/operators maintain an email alias that exists for the sole purpose of receiving disclosure notices from parties reporting noted security flaws on the vendor/operator’s web site.

Further, said email alias should be monitored by individuals with an understanding of web application security issues and business logic flaws, while maintaining a close working relationship with the site developers and operations engineers. This relationship should allow for the quick escalation of reported issues for mitigation and remediation.
Examples of such email alias might include:

Too often vendors and web site operators fail to manage the proper intake and escalation of reported security flaws, leading to lapses in web application security for days, weeks, and even months.

Report Security Flaws will provide resources and guidance for vendors and site operators facing such challenges, with the hope of improving internet security posture for vendor/operators and consumers alike.

If you, dear reader, have tried to no avail to drive a site operator/web vendor/cloud provider to fix flaws, and received no reply, let us know. ReportSecurityFlaws intends to serve as a public motivator to close such gaps and promote improved vendor response, complete with standards.

NOTE: ReportSecurityFlaws is not intended to out vendors who fail to fix. Rather, it is to use all means necessary to ensure they do fix, and promote better standards and practices.

With standards in mind, I've been participating in discussions regarding ISO/IEC 29147, which will hopefully be embraced globally as the ISO standard Security
techniques – Responsible vulnerability disclosure
We also believe there is an opportunity for the PCI Council to incorporate stringent disclosure practices in the PCI DSS.

Disclosure can and must be handled properly. This week's reading included a remarkable and detailed public incident report from Apache to the compromise they'd suffered the week prior. This kind of transparent, open response does, as they clearly state, make the internet a better place. Well done, Apache.

Stay tuned for more....
Cheers. | digg | Submit to Slashdot

Please support the Open Security Foundation (OSVDB)


Anonymous said...

Is there a typo in this line?

"We also believe there is an opportunity for the PCI Council to incorporate string disclosure practices in the PCI DSS."

Was "string" supposed to be "stringent"?

Russ McRee said...

Indeed, it is a typo (now corrected). Thank you.

Moving blog to

toolsmith and HolisticInfoSec have moved. I've decided to consolidate all content on one platform, namely an R markdown blogdown sit...