Conditional flows

You can execute and compare responses from multiple API calls in YAML probes

This instruction is required when you want to fire multiple API calls and compare their responses. A simple use case is probing for Username enumeration vulnerability using password-reset endpoint.

To configure Akto probe YAML for multiple requests, set type: true

execute: 
  type: multiple    # this means multiple requests are configured
  requests: 
   - req: 
     ...
   - req:
     ...  

Instructions available -

For each request, you can now use

• validate block to test for a certain response

• success to specify which request to execute next if validate block returns true

• failure to specify which request to execute next if validate block returns false

• Anywhere, you can set success or failure as vulnerable (to mark a vulnerability) or exit

• Requests are named as x1 x2 etc. automatically. You can use this to jump to a node on success or failure . You can also use them in data operators (eg response payload of x1 should have length > 80 characters)

If success or failure isn't specified, we jump to the next request. If it's the last request, then we jump to the validate block of the probe.

Example - Try a wrong password. If API returns 4xx, then try with wrong usernames

execute:
  type: multiple
  requests:
  - req:
    - modify_body_param:
        password: IncorrectPassword1!   
    - validate:
        response_code:
          gte: 400
          lt: 410
    - success: x2 #move to next request 
    - failure: exit #exit if response is not 4xx
  - req: ....  

Next step -

try with invalid usernames, and check if the response string is different. If they are different, then the application is vulnerable to username enumeration.

  - req:
    - modify_body_param:
        username: [e-mail]
    - validate:    # check if response payload for incorrect user is same as incorrect password
      - response_payload:
          neq: "${x1.response.body}" #x1.response.body refers to the response body of 1st request
    - success: vulnerable #if response payloads differ, then mark it as a vuln.

Complete YAML -

wordLists:
  dummyUsernames:
    - dummy
    - "[e-mail]"
execute:
  type: multiple
  requests:
  - req:
    - modify_body_param:
        password: Default1!
    - validate:
        response_code:
          gte: 400
          lt: 410
    - success: x2
    - failure: exit
  - req:
    - modify_body_param:
        username: ${dummyUsernames}
    - validate:
      - response_payload:
          neq: ${x1.response.body}
    - success: vulnerable