Document Security Redux

Brian Krebs’ Security Fix blog at the Washington Post discusses document security. Anyone who has been a consultant has likely been a victim of information leakage associated with metadata. Most of the time, this is harmless, though possibly embarrassing. Other times, as with some of the incidents Krebs describes, it is worse. I wrote about 29 specific content attributes back in May. Bitform, the company that provided the list, recently released a study of Fortune 100 websites with over 8,000 Microsoft Office files on them. Here are the results:

Target Element

Occurrence Rate

Files Affected

Audio and Video Paths

0.4%

36

Author History

contains paths

contains network share names

46.4%

36.7%

14.4%

3733

2950

1158

Comments

2.1%

165

Content Properties

99.8%

8020

Custom Properties

5.5%

446

Database Queries

0.0%

1

Embedded Objects

24.8%

1994

Encryption

1.1%

92

Fast Save Data

10.1%

813

Hidden Cells

3.9%

315

Hidden Slides

1.9%

151

Hidden Text

0.9%

76

Linked Objects

0.1%

11

Macros and Code

5.1%

409

Office GUID Property

17.2%

1386

Outlook Properties

17.1%

1378

Presentation Notes

13.6%

1093

Printer Information

contains network share names

30.9%

18.0%

2480

1447

Routing Slips

0.0%

0

Scenario Comments

0.0%

0

Sensitive Hyperlinks

0.4%

29

Sensitive Include Fields

0.3%

22

Statistic Properties

99.9%

8028

Summary Properties

99.3%

7978

Template Name

7.4%

592

Tracked Changes

6.5%

521

User Names

98.9%

7950

Versions

0.0%

4

Weak Protections

3.5%

278

For detailed descriptions of each of these target elements, get the paper. Btw, the level of diligence assigned to this information is virtually nil – it is taken at face value. Given that it always appears that public incidents involve the government and/or politics, I wouldn’t put it past political operatives to insert bogus information to harm a competitor or foe.

2 comments for “Document Security Redux

  1. ken
    December 9, 2005 at 3:45 pm

    Here at Workshare, we run a public benefit site at http://www.metadatarisk.org which catalogs all major document leaks of this type. Those who are interested may find it to be a useful resource.
    Ken Rutsky

  2. June 5, 2006 at 3:10 am

    Information is leaked when you forget to make clean of digital data physically clean of metadata. These are simple precaution that is necessary steps. For more information you can see http://www.PinionSoftware.com

Comments are closed.