I am currently seeking a very bright individual to write a piece of code, preferably in .net on windows but theoretically any language on Windows or Linux would be fine, which can reliably defeat an image-based CAPTCHA on a website, for help with a security study project.
## Deliverables
App may run client side on windows (.NET) or server side on Windows or Linux with a client-accessible interface and should be able to meet the following functional targets:
1. Select fields one at a time in sequence from a provided sample delimited txt file
2. Build an HTTP query string incorporating the field value (basic logic provided)
3. Perform HTTP query of a specified website
4. Parse a specified result value from the HTML Output to prove success (logic provided)
5. Write the parsed numeric value to a designated column in the appropriate row of txt file, while preserving file format (field names, delimiting, carriage returns / end of line)
6. Be able to perform this transaction at high speed and in sequence on any large txt file having the same format
7. Circumvent or minimize the effect of volume-triggered image-based CAPTCHA routines to meet a target of 20 or more transactions per minute.
8. Be able to handle delimited txt files with up to 1 million records
9. Provide a command line or graphical interface which can be used to initiate the process and run through a file specified through a command line argument or GUI selector.
Definition of Transaction: A Transaction is defined as one full loop of selecting the record, querying the website, defeating a CAPTCHA if presented, parsing the data and writing the result to the txt file.
Note: Website will not always present a CAPTCHA, it is triggered by a variety of factors including volume, usage pattern etc.
You will need to provide your own development and test environment.
Complete source code must be provided. Code must be commented extensively.
Code must be delivered with compiled version and installation instructions along with an installer OR full installation support from the developer (no remote access) which creates a fully working copy on my own test box in the US, and application must meet performance targets therein
Test for functionality:
------------------------
Runs on server or client side test system
Minimum 20 transactions per minute
Processes 100,000 records successfully
Guaranteed to process 1 million records
We estimate that this is probably 10 hours of work.
Links:
<[login to view URL]>
## Platform
Windows XP (all service packs) or .net server / windows server 2003,
if linux, only server side permissible, no linux GUI dependencies
if web interface enabled internet explorer 6.0