This site is the archived OWASP Foundation Wiki and is no longer accepting Account Requests.
To view the new OWASP Foundation website, please visit https://owasp.org
AppSensor DetectionPoints
About This Document
These detection points are part of the OWASP AppSensor project which advocates bringing intelligent intrusion detection inside the application. These detection points can be used to identify a malicious user that is probing for vulnerabilities or weaknesses within your application.
Read more about why application logging is the way to go.
- 1 About This Document
- 2 Detection Points
- 2.1 RequestException
- 2.1.1 RE1: Unexpected HTTP Command
- 2.1.2 RE2: Attempt to Invoke Unsupported HTTP Method
- 2.1.3 RE3: GET When Expecting POST
- 2.1.4 RE4: POST When Expecting GET
- 2.1.5 RE5: Additional/Duplicated Data in Request
- 2.1.6 RE6: Data Missing from Request
- 2.1.7 RE7: Unexpected Quantity of Characters in Parameter
- 2.1.8 RE8: Unexpected Type of Characters in Parameter
- 2.2 AuthenticationException
- 2.2.1 AE1: Use of Multiple Usernames
- 2.2.2 AE2: Multiple Failed Passwords
- 2.2.3 AE3: High Rate of Login Attempts
- 2.2.4 AE4: Unexpected Quantity of Characters in Username
- 2.2.5 AE5: Unexpected Quantity of Characters in Password
- 2.2.6 AE6: Unexpected Type of Character in Username
- 2.2.7 AE7: Unexpected Type of Character in Password
- 2.2.8 AE8: Providing Only the Username
- 2.2.9 AE9: Providing Only the Password
- 2.2.10 AE10: Additional POST Variable
- 2.2.11 AE11: Missing POST Variable
- 2.2.12 AE12: Utilization of Common Usernames
- 2.3 SessionException
- 2.4 AccessControlException
- 2.5 InputException
- 2.6 EncodingException
- 2.7 CommandInjectionException
- 2.8 FileIOException
- 2.9 Honey Trap
- 2.10 UserTrendException
- 2.11 SystemTrendException
- 2.12 Reputation
- 2.1 RequestException
Detection Points
RequestException
RE1: Unexpected HTTP Command
| ID |
RE1 |
| Title |
Unexpected HTTP Command |
| Category |
RequestException |
| Description |
An HTTP request is received which contains unexpected/disallowed commands. |
| Considerations |
A list of accepted commands should be generated (i.e. GET and POST) and all other HTTP commands should generate an event. See HTTP/1.1: Method Definitions. |
| Tuning |
Browsers and proxies using the HEAD method to check whether the content of a file has changed. |
| Examples |
Instead of a GET or POST request, the user sends a TRACE request to the application. Cross references:
|
| Code |
RE2: Attempt to Invoke Unsupported HTTP Method
| ID |
RE2 |
| Title |
Attempt to Invoke Unsupported HTTP Method |
| Category |
RequestException |
| Description |
An HTTP request is received which contains a non-existent HTTP command (does not match anything in this list: HEAD, GET, POST, PUT, DELETE, TRACE, OPTIONS, CONNECT). |
| Considerations |
- |
| Tuning |
- |
| Examples |
Instead of a GET or POST request, the user sends a TEST request to the application (TEST is not a valid HTTP request method). Cross references:
|
| Code |
RE3: GET When Expecting POST
| ID |
RE3 |
| Title |
GET When Expecting POST |
| Category |
RequestException |
| Description |
A page which is expecting only POST requests, is requested by HTTP method GET. |
| Considerations |
Some pages may be designed to receive both GET and POST requests. |
| Tuning |
Some resources may allow both GET and POST methods e.g. an edit form may be hyperlinked using a parameter value defining the record to be edited, but the form is submitted by POST to itself. Users may bookmark a page that is the result of a POST and return to it at a later date. |
| Examples |
The user sends a GET request to a page which has only been used for POSTs. |
| Code |
RE4: POST When Expecting GET
| ID |
RE4 |
| Title |
POST When Expecting GET |
| Category |
RequestException |
| Description |
A page which is expecting only GET requests, receives a POST. |
| Considerations |
- |
| Tuning |
(same as RE3) |
| Examples |
The user utilizes a proxy tool to build a custom POST request and sends it to a page which has been accessed by GET requests. |
| Code |
RE5: Additional/Duplicated Data in Request
| ID |
RE5 |
| Title |
Additional/Duplicated Data in Request |
| Category |
RequestException |
| Description |
Additional unexpected parameters or HTTP headers, or duplicates, are received with the request. |
| Considerations |
Additional parameters may be an attempt to override values or to exploit unexposed functionality. Duplicated parameters may be an indication of attempted HTTP parameter pollution. Beware of firing this detector when additional cookies, not used by the application, are found (as opposed to duplicated cookies) since these may relate to third-party code (e.g. advertisements, analytics) or some other application. Note that extra HTTP headers may be added by intermediate proxies, and unless the network configuration is fixed (an internal network perhaps), additional headers cannot be controlled and thus cannot be used to infer existence of a potential attacker. |
| Tuning |
Links from third party sites/services may included additional parameters (e.g. from search engines, from advertisements). Additional cookies headers may be added by other applications or by third parties such as advertisers, and there may be very little control over these. Additional HTTP headers may be added by intermediate network devices (e.g. for traffic management). |
| Examples |
Example 1: Additional form or URL parameters submitted with request (e.g. debug=1, servervariable=2000). Example 2: A parameter is defined more than once in the URL Query String. Example 3: An HTTP header is duplicated. Example 4: An additional HTTP header is found. Example 5: A URL path parameter with the same name as a form parameter is sent with the request. Cross references:
|
| Code |
RE6: Data Missing from Request
| ID |
RE6 |
| Title |
Data Missing from Request |
| Category |
RequestException |
| Description |
Expected parameters or HTTP headers are missing from the request. |
| Considerations |
Bookmarking and use of a browser's "back button" can lead to requests without the expected parameters. |
| Tuning |
A bookmarked page may be missing the required POST parameters (see also RE3). Users may choose to send a blank or different User Agent header value. |
| Examples |
Example 1: A page is requested without any of the required form parameters. Example 2: The HTTP-Accept header is not present in a request. |
| Code |
RE7: Unexpected Quantity of Characters in Parameter
| ID |
RE7 |
| Title |
Unexpected Quantity of Characters in Parameter |
| Category |
RequestException |
| Description |
The user provides a parameter value with a large number of characters. |
| Considerations |
- |
| Tuning |
If the input field does not have client-side validation and/or MAXLENGTH attributes, a user might inadvertently copy in some text that is longer than expected. |
| Examples |
Example 1: The user submits a form field with more characters than the form's maxlength attribute and client-side validation would allow. Example 2: The user submits data in a form's hidden field which is longer than expected. Cross references:
|
| Code |
RE8: Unexpected Type of Characters in Parameter
| ID |
RE8 |
| Title |
Unexpected Type of Characters in Parameter |
| Category |
RequestException |
| Description |
The user provides a parameter value containing characters outwith the expected range. |
| Considerations |
- |
| Tuning |
Text fields may include text from copy and paste operations that contain illegal characters. |
| Examples |
Example 1: The user sends an HTTP header containing a line break character. Example 2: The user sends a URL parameter value that contains ASCII characters below 20 or above 7E. Cross references:
|
| Code |
AuthenticationException
AE1: Use of Multiple Usernames
| ID |
AE1 |
| Title |
Use of Multiple Usernames |
| Category |
AuthenticationException |
| Description |
Multiple usernames are attempted when logging into the application. |
| Considerations |
The assignment of login attempts to a user can be based on a sessionID given to the user when they first visit the website. Correlating based on IP address is difficult since multiple users could be using the site from the same IP address (e.g. corporate NAT). |
| Tuning |
- |
| Examples |
User first tries username 'bob', then username 'sue', then 'steve', etc. |
| Code |
AE2: Multiple Failed Passwords
| ID |
AE2 |
| Title |
Multiple Failed Passwords |
| Category |
AuthenticationException |
| Description |
For a single username, multiple bad passwords, or other authentication credentials, are entered. |
| Considerations |
See Popularity is Everything section 4 - Attack-Detection Scenarios for ideas about tracking use of unsuccessful passwords and looking whether these are used against multiple accounts. |
| Tuning |
A users providing the same wrong password more than once may be different to different wrong passwords. See Account Lockout, Bill Cheswick, Episode 76, OWASP Podcast, September 22, 2010. |
| Examples |
Example 1: User tries username:password combination of 'user:pass1', 'user:pass2', 'user:pass3', etc. Example 2: Multiple failed PINs are attempted for the same customer account. Example 3: In an online banking application, several invalid mobile authentication codes, transaction verification codes or transaction authentication numbers are submitted. Example 4: A user provides the correct password, but repeatedly fails to provide the required second password correctly. |
| Code |
AE3: High Rate of Login Attempts
| ID |
AE3 |
| Title |
High Rate of Login Attempts |
| Category |
AuthenticationException |
| Description |
The rate of login attempts becomes too high (possibly indicating an automated login attack). |
| Considerations |
The threshold should relate to a limit and period appropriate to the application (e.g. total number in a second or minute or hour). |
| Tuning |
- |
| Examples |
User sends the following login attempts within 1 second - 'user1:pass1', 'user1:pass2', 'user2:pass3', 'user2:pass4'. |
| Code |
AE4: Unexpected Quantity of Characters in Username
| ID |
AE4 |
| Title |
Unexpected Quantity of Characters in Username |
| Category |
AuthenticationException |
| Description |
The user provides a username with a large number of characters. |
| Considerations |
- |
| Tuning |
(same as RE7) |
| Examples |
The user sends a username that is 200 characters long when 6-8 are expected. |
| Code |
AE5: Unexpected Quantity of Characters in Password
| ID |
AE5 |
| Title |
Unexpected Quantity of Characters in Password |
| Category |
AuthenticationException |
| Description |
The user provides a password with a large number of characters. |
| Considerations |
Higher limits may be required for sites which allow users to have pass phrases. |
| Tuning |
(same as RE7) |
| Examples |
Example 1: The user sends a password that is 200 characters long, when 5-20 are expected. Example 2: The user sends a PIN of 30 characters. |
| Code |
AE6: Unexpected Type of Character in Username
| ID |
AE6 |
| Title |
Unexpected Type of Character in Username |
| Category |
AuthenticationException |
| Description |
The user provides a username which contains characters outwith the expected range. |
| Considerations |
Any characters below hex value 20 or above 7E are often considered illegal (decimal values of below 32 or above 126). |
| Tuning |
Users may be confused between a username, customer identification code and their account number, or even between offline and online identifiers. Mis-typing might add a character like "]" or "#" if these are adjacent to the ENTER/CR key. Whitespace may be appended to values when copied from a spreadsheet cell (e.g. a line feed character when cell values are copied and pasted from Excel). A password may be put in the username field by accident. |
| Examples |
The user sends a username that contains ASCII non-printable characters such as the NULL byte. |
| Code |
AE7: Unexpected Type of Character in Password
| ID |
AE7 |
| Title |
Unexpected Type of Character in Password |
| Category |
AuthenticationException |
| Description |
The user provides a password containing characters outwith the expected range. |
| Considerations |
Examples include null byte, and characters which need the ALT key to be used. |
| Tuning |
(same as AE6) |
| Examples |
The user sends a password that contains ASCII characters below 20 or above 7E. |
| Code |
AE8: Providing Only the Username
| ID |
AE8 |
| Title |
Providing Only the Username |
| Category |
AuthenticationException |
| Description |
The user submits a POST request which only contains the username variable. The password variable has been removed. |
| Considerations |
This is different from only providing the username in the login form since in that case the password variable would be present and empty. |
| Tuning |
- |
| Examples |
The user utilizes a proxy tool to remove the password variable from the submitted POST request. |
| Code |
AE9: Providing Only the Password
| ID |
AE9 |
| Title |
Providing Only the Password |
| Category |
AuthenticationException |
| Description |
The user submits a POST request which only contains the password variable. The username variable has been removed. |
| Considerations |
This is different from only providing the password in the login form since in that case the username variable would be present and empty. |
| Tuning |
- |
| Examples |
The user utilizes a proxy tool to remove the username variable from the submitted POST request. |
| Code |
AE10: Additional POST Variable
| ID |
AE10 |
| Title |
Additional POST Variable |
| Category |
AuthenticationException |
| Description |
Additional, unexpected POST variables are received during an authentication request. |
| Considerations |
- |
| Tuning |
(same as RE5) |
| Examples |
The user utilizes a proxy tool to add the POST variable of 'admin=true' to the request. |
| Code |
AE11: Missing POST Variable
| ID |
AE11 |
| Title |
Missing POST Variables |
| Category |
AuthenticationException |
| Description |
Expected POST variables are not present within the submitted authentication request. |
| Considerations |
- |
| Tuning |
(same as RE6) |
| Examples |
The user utilizes a proxy tool to remove an additional POST variable, such as 'guest=true', from the POST request. |
| Code |
AE12: Utilization of Common Usernames
| ID |
AE12 |
| Title |
Utilization of Common Usernames |
| Category |
AuthenticationException |
| Description |
Common dictionary usernames are used to attempt to log into the application. |
| Considerations |
- |
| Tuning |
Common usernames might be allowed during self-registration or when editing account details. |
| Examples |
Log in attempted with usernames "administrator", then "admin", then "test" |
| Code |
SessionException
SE1: Modifying Existing Cookie
| ID |
SE1 |
| Title |
Modifying Existing Cookie |
| Category |
SessionException |
| Description |
A request is received containing a cookie with a modified value. |
| Considerations |
This could be determined if the cookie is modified to an illegal value. |
| Tuning |
In a poorly designed application, the length of the cookie value, or the combined size of all the cookies, might possibly exceed that which is supported. |
| Examples |
Example 1: The user utilizes a proxy tool to change the encrypted cookie to an alternative value which does not properly decode within the application. Example 2: The user modifies an unencrypted cookie and sets an illegal value for a particular variable. |
| Code |
SE2: Adding New Cookie
| ID |
SE2 |
| Title |
Adding New Cookie |
| Category |
SessionException |
| Description |
A request is received which contains additional cookies that are not expected by the application. |
| Considerations |
A session cookie existing when it should not (e.g. prior to authentication) is probably indicative of an attack. But cookies may also be set by third party sites which get send with the request - these may be harmless. Also consider what other applications exist on sub-domains (e.g. www.example.com, extranet.example.com and sales.example.com) which may also be setting cookies. |
| Tuning |
- |
| Examples |
The user utilizes a proxy tool to add cookies to the request. |
| Code |
SE3: Deleting Existing Cookie
| ID |
SE3 |
| Title |
Deleting Existing Cookie |
| Category |
SessionException |
| Description |
A request is received which does not contain the expected cookies. |
| Considerations |
The user may have bookmarked a page they had visited during a previous authenticated session. |
| Tuning |
In a poorly designed application, the number of cookies might exceed the allowed number supported by the user's browser. |
| Examples |
The user utilizes a proxy tool to remove cookies or portions of cookies from a request. |
| Code |
SE4: Substituting Another User's Valid Session ID or Cookie
| ID |
SE4 |
| Title |
Substituting Another User's Valid Session ID or Cookie |
| Category |
SessionException |
| Description |
A request is received which contains cookie data that is clearly from another user or another session. |
| Considerations |
- |
| Tuning |
A mis-configured proxy might send the same session ID or cookie for all users. |
| Examples |
The user utilizes a proxy tool to substitute valid data from another user or session into the cookie. An example would be changing some sort of identification number within the cookie. |
| Code |
SE5: Source Location Changes During Session
| ID |
SE5 |
| Title |
Source Location Changes During Session |
| Category |
SessionException |
| Description |
Valid requests, containing valid session credentials, are received from multiple source locations indicating a possible session hijacking attack. |
| Considerations |
A full IP address may not be constant for some users during normal use. Enforcing single fixed IP addresses for each session in an intranet application may be valid. However, if the application is accessible over public networks, changing IP address cannot be excluded and it may be more useful to consider fixing just part of the IP address, or looking for more significant changes such as when the user's IP address geo-location region or country changes (see Autonomous System Number (ASN) and Detecting Malice with ModSecurity: GeoLocation Data). Note: source port number should not be used in checks since this usually changes very frequently. |
| Tuning |
If the full IP address is used for this, it may change slightly from request to request by the same user. |
| Examples |
Example 1: User A's session is compromised and User B begins using the account. The requests originating from User B will possibly contain a different source IP address the User A. The source IP addresses could be the same if both users where behind the same NAT. Example 2: An application at a fixed server location, which calls web services, changes IP address unexpectedly. |
| Code |
SE6: Change of User Agent Mid Session
| ID |
SE6 |
| Title |
Change of User Agent Mid Session |
| Category |
SessionException |
| Description |
The User-Agent value of the header changes during a session. This indicates a different browser is now being used. Although this value is under the control of the sender, a change in this may indicates that the session has been compromised and is being used another individual. This will likely not be the case that the user has simply copied and pasted the URL from one browser to another on the same system because this action would not copy over the appropriate session identifiers. |
| Considerations |
Optionally also include other HTTP headers in this check. For example, the Accept-Encoding and Accept-Language headers do not normally change and could be concatenated with the User-Agent and hashed to created an identifier. The ideas described in Panopticlick and Javascript Browser Fingerprinting can also be used to fingerprint a particular client system but require the use of client-side code. Application owners should check the legality of collecting data, and whether it is considered "personal data" which may have additional constraints in some jurisdictions. |
| Tuning |
- |
| Examples |
Mid session, the User Agent changes from Firefox to Internet Explorer. |
| Code |
AccessControlException
ACE1: Modifying URL Argument Within a GET for Direct Object Access Attempt
| ID |
ACE1 |
| Title |
Modifying URL Argument Within a GET for Direct Object Access Attempt |
| Category |
AccessControlException |
| Description |
The application is designed to use an identifier for a particular object, such as using categoryID=4 or user=guest within the URL. A user modifies this value in an attempt to access unauthorized information. This exception should be thrown anytime the identifier received from the user is not authorized due to the identifier being non-existent or the identifier not authorized for that user. |
| Considerations |
- |
| Tuning |
Bookmarking , truncation, and mistyping issues could lead to some access control exceptions. |
| Examples |
The user modifies the following URL from example.com/viewpage?page=1&user=guest to example.com/viewpage?page=22&user=admin |
| Code |
ACE2: Modifying Parameter Within A POST for Direct Object Access Attempt
| ID |
ACE2 |
| Title |
Modifying Parameter Within A POST for Direct Object Access Attempt |
| Category |
AccessControlException |
| Description |
The value of a non-free text html form element (i.e. drop down box, radio button) is modified to an illegal value. The value either does not exist or is not authorized for the user. |
| Considerations |
- |
| Tuning |
(same as ACE1 for bookmarking) |
| Examples |
The user utilizes a proxy tool to intercept a POST request and changes the submitted value to a value that was not available through the normal display. For example, the user encounters a dropdown box containing the numbers 1 through 10. The user selects 5 and then intercepts the request to change the submitted value to 100. |
| Code |
ACE3: Force Browsing Attempt
| ID |
ACE3 |
| Title |
Force Browsing Attempt |
| Category |
AccessControlException |
| Description |
An authenticated or unauthenticated user sends a request for a non-existent resource (e.g. page, directory listing, image, file, etc), or a resource that is not authorized for that user. |
| Considerations |
- |
| Tuning |
Requests for non-existent resources may occur for many reasons such as Benign Unexpected URLs - Part 1 - Missing (404 Not Found Error) Files |
| Examples |
Example 1: The user is authenticated and requests site.com/PageThatDoesNotExist. Example 2: The user is authenticated and requests a video they are not authorized to download/view. Example 3: An unauthenticated user (perhaps with a session ID) requests a listing of a directory detailed in the site's robots.txt file. |
| Code |
ACE4: Evading Presentation Access Control Through Custom POST
| ID |
ACE4 |
| Title |
Evading Presentation Access Control Through Custom POST |
| Category |
AccessControlException |
| Description |
A POST request is received which is not authorized for the current user and the user could not have performed this action without crafting a custom POST request. |
| Considerations |
This situation is most likely to occur when presentation layer access controls are in place and have removed the user's ability to initiate the action through the presentation of the application. An attacker may be aware of the functionality and attempt to bypass this presentation layer access control by crafting their own custom message and sending this in an attempt to execute the functionality. |
| Tuning |
- |
| Examples |
The application contains the ability for an administrator to delete a user. This method is normally invoked by entering the username and submitting to https://oursite/deleteuser Presentation layer access controls ensure the delete user form is not displayed to non-administrator users. A malicious user has access to a non-administrator account and is aware of the delete user functionality. The malicious user sends a custom crafted POST message to https://oursite/deleteuser in an attempt to execute the delete user method. |
| Code |
InputException
IE1: Cross Site Scripting Attempt
| ID |
IE1 |
| Title |
Cross Site Scripting Attempt |
| Category |
InputException |
| Description |
The HTTP request contains common XSS attacks which are often used by attackers probing for XSS vulnerabilities. |
| Considerations |
Detection should be configured to test all GET and POST values as well as all header names and values for the following values. |
| Tuning |
There are many patterns which could be XSS but may also be normal user input to a free text field e.g. "Press the 'drop' button" if a pattern were looking for a single quotation mark followed by SQL commands like DROP, INSERT, UPDATE and DELETE. Applications that are used to discuss or share information about programming, software development and security may want to allow such free text input, provided it is encoded/escaped correctly. |
| Examples |
The user utilizes a proxy tool to add an XSS attack to the header value and the 'displayname' POST variable. The header value could be displayed to an admin viewing log files and the 'displayname' POST variable may be stored in the application and displayed to other users. Note, the following XSS attacks would only be used by an attacker to probe for vulnerability. An actual XSS attack would be customized by the attacker. <script>alert(document.cookie);</script> <script>alert();</script> alert(String.fromCharCode(88,83,83)) <IMG SRC="javascript:alert('XSS');"> <IMG SRC=javascript:alert('XSS')> <IMG SRC=javascript:alert("XSS")> <BODY ONLOAD=alert('XSS')> Cross references:
|
| Code |
IE2: Violation Of Implemented White Lists
| ID |
IE2 |
| Title |
Violation Of Implemented White Lists |
| Category |
InputException |
| Description |
The application receives user-supplied data that violates an established white list validation. |
| Considerations |
See AC3 (Force Browsing Attempts) about requests for non-existent/unauthorised (i.e. not white listed) URLs. |
| Tuning |
(same as IE1) |
| Examples |
The user submits data that is not correct for the particular field. This may not be attack data necessarily, but repeated violations could be an attempt by the attacker to determine how an application works or to discover a flaw. |
| Code |
IE3: Violation Of Implemented Black Lists
| ID |
IE3 |
| Title |
Violation Of Implemented Black Lists |
| Category |
InputException |
| Description |
The application receives user-supplied data that violates an established black list validation. |
| Considerations |
This may not be attack data necessarily, but repeated violations could be an attempt by the attacker to determine how an application works or to discover a flaw or to exploit a flaw. This black list approach suffers from the potential for greater false positives than IE2 above, and cannot be used to identify all potential malicious data. |
| Tuning |
(same as IE1) |
| Examples |
Example 1: URL in comment field identified as suspected phishing and malware pages using Google Safe Browsing API). Example 2: Parameter value matches a known SQL injection pattern. Example 3: Parameter value matches a known XSS pattern. |
| Code |
IE4: Violation of Input Data Integrity
| ID |
IE4 |
| Title |
Violation of Input Data Integrity |
| Category |
InputException |
| Description |
The application receives HTTP header or body parameter values which have been tampered with when no change should have occurred. |
| Considerations |
- |
| Tuning |
This detection point should only be used with parameters that cannot be altered by accident. Input types text and textarea would not normally be suitable, even if the elements are disabled in the browser. Be wary of assuming JavaScript will prevent modification of form elements in all conditions. |
| Examples |
Example 1: Hidden form field modified by client. Example 2: Select list value submitted in response, not sent by server as an available option value. Example 3: Cookie set by server has been manipulated by the client. Example 4: Cookie created by client instead of by the server. |
| Code |
IE5: Violation of Stored Business Data Integrity
| ID |
IE5 |
| Title |
Violation of Stored Business Data Integrity |
| Category |
InputException |
| Description |
User's input leads to violation of data integrity. |
| Considerations |
- |
| Tuning |
- |
| Examples |
Example 1: A user's action leads to a system integrity error when writing to, or updating, a database. Example 2: Business rule checks detect that a user's action is not compatible. Example 3: Data accuracy checking detects duplicate records for a user. Example 4: User input leads to an unexpected file change (e.g. .htaccess file overwritten, page template changed). Example 5: User's request leads to a new, unexpected, outbound network connection being made (e.g. mail sent to an SMTP server, files downloaded from a FTP server). |
| Code |
IE6: Violation of Security Log Integrity
| ID |
IE6 |
| Title |
Violation of Security Log Integrity |
| Category |
InputException |
| Description |
Security or audit log tampering detected. |
| Considerations |
AppSensor may rely on the accuracy of "log" data to make decisions when thresholds are reached. This detector aims to detect the insertion of forged entries, corruption of logs, unauthorised deletion of and changes to records. See also: |
| Tuning |
- |
| Examples |
Example 1: Special characters embedded in logged data about a user's activity cause the data to overwrite a previous log entry. Example 2: Log file integrity is broken by modification to an existing log entry. |
| Code |
EncodingException
EE1: Double Encoded Character
| ID |
EE1 |
| Title |
Double Encoded Characters |
| Category |
EncodingException |
| Description |
An HTTP request is received which contains one or more double encoded values. |
| Considerations |
- |
| Tuning |
Data supplied by other party systems may have encoding issues. |
| Examples |
The user sends encodes the % symbol to %25 and appends 3C. The user is sending %253C which may be interpreted by the application as %3C which is actually <. |
| Code |
EE2: Unexpected Encoding Used
| ID |
EE2 |
| Title |
Unexpected Encoding Used |
| Category |
EncodingException |
| Description |
An HTTP request is received which contains values that have encoded in an unexpected format. |
| Considerations |
- |
| Tuning |
(same as EE1) |
| Examples |
The user encodes an attack such as alert(document.cookie) into the UTF-7 format and sends this data the application. This could bypass validation filters and be rendered to a user in certain situations. Cross references:
|
| Code |
CommandInjectionException
CIE1: Blacklist Inspection for Common SQL Injection Values
| ID |
CIE1 |
| Title |
Blacklist Inspection for Common SQL Injection Values |
| Category |
CommandInjectionException |
| Description |
A request is received which contains common SQL injection attack attempts. |
| Considerations |
The point of this detection is not to detect all variations of a SQL injection attack, but to detect the common probes which an attacker or tool might use to determine if a SQL injection vulnerability is present. Unless the site contains some sort of message board for discussing SQL injection, there is little reason that the SQL injection examples should ever be received from a user request. |
| Tuning |
(same as IE1) |
| Examples |
The user sends a request and modifies a URL parameter from category = 5 to category = 5' OR '1' = '1 in an attempt to perform an SQL injection attack. The user could perform similar attacks by modifying POST variables or even the request headers to contain SQL injection attacks. ' OR '1'='1 ' OR 'a'='a ' OR 1=1-- xp_cmdshell UNION JOIN Cross references:
|
| Code |
CIE2: Detect Abnormal Quantity of Returned Records
| ID |
CIE2 |
| Title |
Detect Abnormal Quantity of Returned Records |
| Category |
CommandInjectionException |
| Description |
A database query is executed which returns more records than expected. |
| Considerations |
- |
| Tuning |
- |
| Examples |
Example 1: A query of a non-SQL dataset should only return 1 record but 100 records are returned. Example 2: The application is designed to allow a user to maintain 5 profiles. A user makes a request to view all of their profiles. The database SQL query, which is expected to always return 5 or less results, returns 10,000 records. Something in the application, or user's actions, has caused unauthorized data to be returned. Example 3: Extraction of data from an XML file should only return one matching node, but more than one is returned. |
| Code |
CIE3: Null Byte Character in File Request
| ID |
CIE3 |
| Title |
Null Byte Character in File Request |
| Category |
CommandInjectionException |
| Description |
A request is received to download a file from the server. The filename requested contains the null byte the file name. This is an attempted OS injection attack. |
| Considerations |
- |
| Tuning |
- |
| Examples |
The user modifies the filename of the requested file to download to contain the null byte. The null byte can be added by inserting the hex value %00. |
| Code |
CIE4: Carriage Return or Line Feed Character in File Request
| ID |
CIE4 |
| Title |
Carriage Return or Line Feed Character in File Request |
| Category |
CommandInjectionException |
| Description |
A request is received which contains the carriage return or line feed characters within the POST data or the URL parameters. This is an attempted HTTP split response attack. |
| Considerations |
- |
| Tuning |
- |
| Examples |
The user includes the hex value %0D or %0A in the HTTP request POST data or URL parameters. Cross references:
|
| Code |
FileIOException
FIO1: Detect Large Individual File
| ID |
FIO1 |
| Title |
Detect Large Individual File |
| Category |
FileIOException |
| Description |
A file upload feature detects that a large file has been submitted for upload which exceeds the maximum upload size. |
| Considerations |
- |
| Tuning |
- |
| Examples |
The user attempts to upload a large file to occupy resources or fill up disk space. |
| Code |
FIO2: Detect Large Number of File Uploads
| ID |
FIO2 |
| Title |
Detect Large Number of File Uploads |
| Category |
FileIOException |
| Description |
A user uploads an excessively large number of files. |
| Considerations |
The limit and period used to determine the threshold rate is application dependent, and may also depend on the user's role. |
| Tuning |
- |
| Examples |
A single user attempts to upload multiple small files to occupy resources or fill up disk space. |
| Code |
Honey Trap
HT1: Alteration to Honey Trap Data
| ID |
HT1 |
| Title |
Alteration to Honey Trap Data |
| Category |
HoneyTrap |
| Description |
Fake (not otherwise needed by the application) data sent to the user and returned (e.g. as form, URL, cookie values or in the path or HTTP header) is modified. This is usually combined with making the name or value a tempting item for an attacker to try modifying. |
| Considerations |
Similar techniques can also be used for the creation of accessible CAPTCHA. See also ideas at http://blogs.sans.org/appsecstreetfighter/2009/06/04/my-top-6-honeytokens/ |
| Tuning |
- |
| Examples |
Example 1: Otherwise useless hidden fields, which look like potential vulnerabilities, added to some forms are sent back to the server modified (e.g. <input type="hidden" name="admin" value="false" />). Example 2: An additional URL parameter, which is not used by the application, is modified by the user (e.g. www.example.com/account.jsp?debug=0). Example 3: An additional fake cookie is added and is modified by the user. Example 4: URL rewriting is used and a fake directory name is added; this is modified by the user (e.g. www.example.com/orders/normaluser/display.php). Cross references:
|
| Code |
HT2: Honey Trap Resource Requested
| ID |
HT2 |
| Title |
Honey Trap Resource Requested |
| Category |
HoneyTrap |
| Description |
A purposely leaked resource that has no use in normal application use is requested by a user. |
| Considerations |
Ensure the resource is not linked from normal application content such that a spider or robot might find the resource in any case. |
| Tuning |
- |
| Examples |
Example 1: Page, directory or other resource listed in the application's robots.txt robots exclusion file is requested by the user. Example 2: URL identified only in HTML comments is requested by the user. Example 3: Unexposed server function call included in Flash file source code is requested by the user. |
| Code |
HT3: Honey Trap Data Used
| ID |
HT3 |
| Title |
Honey Trap Data Used |
| Category |
HoneyTrap |
| Description |
Special data sent or accessed by a user. |
| Considerations |
For honey trap data that is detected on egress only, use of outbound content monitoring (e.g. a web application firewall or similar technique) may be helpful. |
| Tuning |
- |
| Examples |
Example 1: Fake user name and password only visible in source HTML code used to attempt to log in to the application (e.g. in HTML comments, in server-side code 'accidentally' delivered to the user). Example 2: A special code number or account name is left in a discussion forum site; this is then used in the application. Example 3: An attempt is made to authenticate with the user name listed in the first row (e.g. ID=1) of the application's database table of Users. Example 4: Data from a fake account record is sent by the server and detected; this record should not normally be accessible by anyone using the application. |
| Code |
UserTrendException
UT1: Irregular Use of Application
| ID |
UT1 |
| Title |
Irregular Use of Application |
| Category |
UserTrendException |
| Description |
The application receives an unusual pattern of requests for the same page or feature from a user. The user may be sending different data combinations or trying to detect errors in the page. |
| Considerations |
- |
| Tuning |
Use of bookmarked URLs and the "back" button might generate out-of-sequence requests. See also related frequency of feature use in UT4. |
| Examples |
Example 1: The user requests a particular page, such as the address update page, numerous times. Example 2: The user requests a page out-of-sequence, such as an intermediate step in a multi-stage form. Example 3: The user only requests dynamic content, and not the associated static files (e.g. images, stylesheets). Example 4: The user sends a slow request/read in an attempt at application denial of service. |
| Code |
UT2: Speed of Application Use
| ID |
UT2 |
| Title |
Speed of Application Use |
| Category |
UserTrendException |
| Description |
The speed of requests from a user indicates that an automated tool is being used to access the site. The use of a tool undertaking a high number of requests quickly may indicate reconnaissance for an attack or attempts to identify vulnerabilities in the site. Slow usage (e.g. between account creation and use) might indicate automated account creation that are then used subsequently for attacks. |
| Considerations |
If enforced inappropriately or too rigorously, this detection point could lead to false positives. |
| Tuning |
Time periods need to be set broadly enough to cater for the normal spread in user behavior. Some users may use automated tools that store passwords securely to populate and submit authentication forms. |
| Examples |
Example 1: The user utilizes an automated tool to request hundreds of pages per minute. Example 2: The user does not log in to the site until a long time after their account is created. Example 3: New (uncached) static content such as images and style sheets associated with each page are not requested in a similar time period as the page. Example 4: A CAPTCHA challenge is responded to more quickly than a human could possibly do. Example 5: The user's clickstream data velocity is too high. |
| Code |
UT3: Frequency of Site Use
| ID |
UT3 |
| Title |
Frequency of Site Use |
| Category |
UserTrendException |
| Description |
Change in how often the site is used by a user |
| Considerations |
- |
| Tuning |
Some users may correctly change their behavior in the frequency of accessing the application. |
| Examples |
The user normally accesses the site once per week, but this changes to many times per day. |
| Code |
UT4: Frequency of Feature Use
| ID |
UT4 |
| Title |
Frequency of Feature Use |
| Category |
UserTrendException |
| Description |
The rate of a user utilizing a particular application feature changes dramatically. |
| Considerations |
- |
| Tuning |
It may be valid for some functionality may be requested repeatedly. For example a real customer placing many orders, a press officer publishing a backlog of press releases, or an administrator populating a staff directory. |
| Examples |
Example 1: The user submits many forum messages in a short period of time. Example 2: The user adds many new friends rapidly. |
| Code |
SystemTrendException
STE1: High Number of Logouts Across The Site
| ID |
STE1 |
| Title |
High Number of Logouts Across The Site |
| Category |
SystemTrendException |
| Description |
A sudden spike in logouts across the application could indicate a XSS and CSRF attack placed within the application which is automatically logging off users. |
| Considerations |
- |
| Tuning |
- |
| Examples |
The hourly usage of the log-off feature of the application suddenly spikes by 500%. |
| Code |
STE2: High Number of Logins Across The Site
| ID |
STE2 |
| Title |
High Number of Logins Across The Site |
| Category |
SystemTrendException |
| Description |
A sudden spike in logins across the application could indicate users being redirected to the site from a phishing email looking to exploit a XSS vulnerability in the site. |
| Considerations |
- |
| Tuning |
- |
| Examples |
The hourly usage of the logon feature of the application suddenly spikes by 1,000%. |
| Code |
STE3: Significant Change in Usage of Same Transaction Across The Site
| ID |
STE3 |
| Title |
Significant Change in Usage of Same Transaction Across The Site |
| Category |
SystemTrendException |
| Description |
A sudden spike in similar activity across numerous users of the application may indicate a phishing attack or CSRF attack against the users; a rapid reduction in activity may indicate users are being redirected elsewhere; a significant change in average transaction value or other quantitative measure may indicate suspicious activity. |
| Considerations |
External events (e.g. a news item) may lead to additional unexpected traffic which is not an attack. |
| Tuning |
A special requirement, situation or event may dramatically change the rate of use of certain transactions. (See also UT4) |
| Examples |
Example 1: The hourly usage of the update email address feature of the application suddenly spikes by 2,000%. Example 2: A website is compromised and users are redirected to a malicious site part-way through a process; the number of successful fully completed transactions drops to nil. Example 3: A number of slow requests/reads are received in an attempt at application denial of service. |
| Code |
Reputation
These reputation detection points could be treated either as:
- like any other detection point contributing to the count of suspicious events, or
- used to alter security logging, or the threshold levels, or associated response actions
The former could lead to a much higher false positive rate.
RP1: Suspicious or Disallowed User Source Location
| ID |
RP1 |
| Title |
Suspicious or Disallowed User Source Location |
| Category |
Reputation |
| Description |
The user is identified as using an IP address associated with a blacklist |
| Considerations |
Suspicious or invalid geo-location, IP addresses or IP address ranges may be identified using a whitelist,internal blacklist, list of Tor nodes (e.g. https://torstat.xenobite.eu/), HTTP blacklist (e.g. http://www.projecthoneypot.org/httpbl.php and Dshield http://www.dshield.org) list of spammers (e.g. Spamhaus http://www.spamhaus.org/) or known botnets (e.g. http://www.shadowserver.org/wiki/). "Suspicious" may also depend upon the type of user e.g. users in the "CMS manager" role should be using an internal network IP address, public users could be from anywhere, customers should only be accessing the application from a particular geographical region, search engine robots should be from a limited range of IP addresses. Take care that "suspicious" does not contribute to greater false positives. |
| Tuning |
The currency and accuracy of needs to be considered when the information is used in AppSensor. The method of challenge and removal of inaccuracies, and the speed of this process, should also be considered. |
| Examples |
Example 1: A user with an external IP address is accessing an internal application, which should not be occurring. Example 2: An authenticated user is accessing the application using a known Tor node, and attack detection thresholds are made more strict. Example 3: An authenticated user is accessing the application from a known trustworthy IP address, and thresholds for certain activity (e.g. input data validation errors) are relaxed slightly. Example 4: The IP address of the payment authentication server, used by the application for credit card authorisation, changes. |
| Code |
RP2: Suspicious External User Behavior
| ID |
RP2 |
| Title |
Suspicious External User Behavior |
| Category |
Reputation |
| Description |
External (to the application) devices and systems (e.g. host and network IDS, file integrity monitoring, disk usage monitoring, anti-malware service, IPS, network firewall, web application firewall, web server logging, XML gateway, database firewall, SIEM) detect anomalous behavior by the user (e.g. session and/or IP address). This information can be used by the application to contribute to its knowledge about a potential attacker. In some cases, the information could be detected by the application itself (e.g. XSS pattern black listing), but may be more effectively identified by the external device, or is not known to the application normally (e.g. requests for missing resources that the web server sees, but does not pass onto the application). |
| Considerations |
The greater the knowledge a device or system has about the application, the greater confidence can be given to evidence of suspicious behaviour. Therefore, for example, attempted SQL injection detected by a web application firewall (WAF) might be given greater weight than information from a network firewall about the IP address. The power of AppSensor is its accuracy and low false positive rate, and the usage of external data should be carefully assessed to ensure it does not contribute to a higher false positive rate. |
| Tuning |
The level of trust in information from the external device/system/organization needs to be considered. |
| Examples |
Example 1: A network IDS has detected suspicious activity by a particular IP address, and this is used to temporarily tighten the attack detection thresholds for requests from all users in the same IP address range. Example 2: An application is using the ModSecurity web application firewall with the Core Rule Set, and utilises the anomaly score data passed forward in the X-WAF-Events and X-WAF-Score HTTP headers (optional rules in modsecurity_crs_49_header_tagging.conf) to adjust the level of application logging for each user. Example 3: Information from an instance of PHPIDS suggests request data may be malicious. |
| Code |
RP3: Suspicious Client-Side Behavior
| ID |
RP3 |
| Title |
RP3 Suspicious Client-Side Behavior |
| Category |
Reputation |
| Description |
The application receives a report of client-side security policy exceptions. |
| Considerations |
Take care this information does not contribute to greater false positives. |
| Tuning |
- |
| Examples |
Example 1: An internal corporate intranet application detects use of a non-standard workstation configuration (e.g. using JavaScript font or plugin detection see SE6). An alert is raised for further investigation. Example 2: An online banking application receives details of suspicious client-side behaviour that would not be expected in normal application use, via a Content Security Policy violation report. The application increases logging for the user, and decreases the monetary limit at which the user's payments require manual authorisation by bank staff. Example 3: The HTTP user agent header value does not agree with other indicators (e.g. using JavaScript detection as in example 1 above). Reference. Example 4: A honeyclient system monitoring the web application reports unexpected behavior in the generated web pages output. Example 5: A third-party monitoring system detects page content that is unauthorised and/or contrary to policy (e.g. structure, included links, HTML validation, inclusion of certain data such as payment card data). Cross references:
|
| Code |
RP4: Change to Environment Threat Level
| ID |
RP4 |
| Title |
Change to Environment Threat Level |
| Category |
Reputation |
| Description |
The general threat level (e.g. general risk of attack from the Internet, or specific targetted attacks against an organisation) is elevated. This could also be used to change response sensitivity due to short-term effects such as application upgrades/patching. |
| Considerations |
This input could be used to alter thresholds for AppSensor responses. |
| Tuning |
The detection point could receive specially crafted input from an attacker, and therefore the information should be considered as untrusted. |
| Examples |
Example 1: A machine-readable threat index is read from a third-party and is used to control security logging levels. Example 2: Business circumstances (e.g. increased attention by activists) raises the suspicion the application may be at increased risk of mis-use, and response thresholds for attack detection are tightened for non-authenticated users. Example 3: The Defense Condition Level (DEFCON) is raised and response thresholds are changed. Example 4: Sensor signal missing. Example 5: External power source disconnected. Example 6: Firmware or software patch signing check failure. |
| Code |
