How to capture HTTP request / response headers with mitmproxy?

Question

I have been able to capture the HTTP(s) traffic from a smartphone and also stored this traffic using mitmdump using the command

mitmdump -w outfile

This seems to dump the HTTP body along with the headers as well. I am interested in capturing only the headers, prefarably as a single csv row (or json string). How can I do that?

Sorites

did you try a filter already ? mitmproxy.org/doc/features/filters.html – Capful

Answer 1

Yet another derived snippet based on previous responses and updated to python3:

def response(flow):
    print("")
    print("="*50)
    #print("FOR: " + flow.request.url)
    print(flow.request.method + " " + flow.request.path + " " + flow.request.http_version)

    print("-"*50 + "request headers:")
    for k, v in flow.request.headers.items():
        print("%-20s: %s" % (k.upper(), v))

    print("-"*50 + "response headers:")
    for k, v in flow.response.headers.items():
        print("%-20s: %s" % (k.upper(), v))
        print("-"*50 + "request headers:")

Command line:

mitmdump -q -v -s parse_headers.py -R http://localhost:9200 -p 30001

Output:

==================================================
GET / HTTP/1.1
--------------------------------------------------request headers:
CONTENT-TYPE        : application/json
ACCEPT              : application/json
USER-AGENT          : Jakarta Commons-HttpClient/3.1
HOST                : localhost
--------------------------------------------------response headers:
CONTENT-TYPE        : application/json; charset=UTF-8
CONTENT-LENGTH      : 327

Popularly

Very nice. I've written up some instructions on how to deploy this and tidied up a few small things in your script: gist.github.com/tomsaleeba/c463550b43eb9c58d8b415523c49f70b – Brakeman

Answer 2

You can extract any header fields you need, e.g., with mitmdump and the flow object (python inline scripts). Inline scripts are documented here: https://mitmproxy.org/doc/scripting/inlinescripts.html

To extract all headers, I used the following command:

$ mitmdump -n -q -s parse_headers.py -r <file>.mitm

The parse_headers.py inline script is as follows:

def response(context, flow):
    request_headers = [{"name": k, "value": v} for k, v in flow.request.headers]
    response_headers = [{"name": k, "value": v} for k, v in flow.response.headers]
    print request_headers
    print response_headers

Immodest

I got no output other than Script error: too many values to unpack Script error: too many values to unpack... :( – Stealth

Answer 3

U was using @rvaneijk, but I was getting the following error:

Script error: too many values to unpack 
Script error: too many values to unpack

I found a solution at 'too many values to unpack', iterating over a dict. key=>string, value=>list and changed the code as follows:

[root@npmjs npmo-server]# cat parse_headers.py
def response(context, flow):
  request_headers = [{"name": k, "value": v} for k, v in flow.request.headers.iteritems()]
  response_headers = [{"name": k, "value": v} for k, v in flow.response.headers.iteritems()]
  print "################################"
  print "FOR: " + flow.request.url
  print flow.request.method + " " + flow.request.path + " " + flow.request.http_version
  print "HTTP REQUEST HEADERS"
  print request_headers
  print "HTTP RESPONSE HEADERS"
  print response_headers
  print ""

The output of this is as follows:

10.137.66.4:63870: clientdisconnect

################################
FOR: http://pe2enpmas300.corp.company.net:8081/csv-stringify
GET /csv-stringify HTTP/1.1
HTTP REQUEST HEADERS
[{'name': 'accept-encoding', 'value': 'gzip'}, {'name': 'authorization', 'value': 'Bearer d2e0770656a9726dfb559ea2ddccff3078dba9a0'}, {'name': 'version', 'value': '2.11.2'}, {'name': 'accept', 'value': 'application/json'}, {'name': 'referer', 'value': 'install restify'}, {'name': 'npm-session', 'value': 'a9a4d805c6392599'}, {'name': 'user-agent', 'value': 'npm/2.11.2 node/v0.10.25 linux x64'}, {'name': 'if-none-match', 'value': 'W/"43fb-8/w7tzRZ9CvawCJo5Uiisg"'}, {'name': 'host', 'value': 'registry-e2e.npmjs.intuit.net'}, {'name': 'Connection', 'value': 'keep-alive'}, {'name': 'X-Forwarded-For', 'value': '10.181.70.43'}]
HTTP RESPONSE HEADERS
[{'name': 'X-Powered-By', 'value': 'Express'}, {'name': 'ETag', 'value': 'W/"43fb-8/w7tzRZ9CvawCJo5Uiisg"'}, {'name': 'Date', 'value': 'Tue, 18 Oct 2016 08:04:45 GMT'}, {'name': 'Connection', 'value': 'keep-alive'}]

You can use Docker as follows:

1. Create the file locally
2. Run the following

Make sure you have read permission on the file.

docker run -ti -p 8080:8080 -v $PWD/parse_headers.py:/tmp/parse_headers.py 
    mitmproxy/mitmproxy mitmdump -s /tmp/parse_headers.py 
    -R http://npmjs.corp.company.net:8081 8080

Stealth