In hrbrmstr/r7snr: r7snr title goes here

r7snr : Tools to work with Rapid7 scans.io Sonar Data

The following functions are implemented:

• snr_parse_response: Parse Sonar HTTP study encoded data response.

Installation

devtools::install_github("hrbrmstr/r7snr")

options(width=120)

Usage

This R package will let you work directly with the gzip'd JSON from Rapid7 scans.io Sonar HTTP studies.

Be warned that these are HUGE files and it's very likely this won't fit into memory on any system.

library(r7snr)

# current verison
packageVersion("r7snr")

library(r7snr)
library(jsonlite)
library(purrr)
library(dplyr)
library(purrr)

For this example, we'll grab the first 10 records from the 2016-07-05 HTTP study. It may report a message of "gzcat: error writing to output: Broken pipe" that we can ignore

system("curl --silent 'https://scans.io/data/rapid7/sonar.http/20160705-http.gz' | gzcat | head -100",
       intern=TRUE) %>%
  map_df(fromJSON) -> http_scan_records

We can take a look a the scan records to find that we have the:

• virtual host name (this will have real virtual host names in forthcoming studies
• the host (both of these are reported by zmap/zgrab)
• data, which is a base 64 encoded version of the HTTP response (up to 32K)
• port, which is the port the grabber connected to
• ip, the IP address the grabber connected to

glimpse(http_scan_records)

We can examine one of them:

str(snr_parse_response(http_scan_records$data[10]))

Or, we can turn them all into a data frame:

map_df(1:nrow(http_scan_records), function(i) {
  x <- http_scan_records[i,]
  resp <- snr_parse_response(x$data)[[1]]
  data_frame(
    vhost=x$vhost, 
    host=x$host,
    port=x$port,
    ip=x$ip,
    status=resp$status,
    version=resp$version,
    body=resp$body,
    headers=list(resp$headers, stringsAsFactors=FALSE)
  )
}) -> parsed_records

Now we have a fairly usable data frame.

glimpse(parsed_records)

We can now see the server types for what we've read in. Since that is not a required header, it may not be there so we have to handle NULL values.

map(parsed_records$headers, "server") %>% 
  map_chr(function(x) ifelse(length(x)>0, x, NA)) %>% 
  table(exclude=FALSE) %>% 
  as.data.frame(stringsAsFactors=FALSE) %>% 
  setNames(c("type", "count")) %>% 
  arrange(desc(count))

If you're going to use this to process the Sonar HTTP studies, it's suggested you will want to use the jqr package to filter the downloaded gzip'd JSON file (which does mean learning the jq filter syntax).

An alternative to using this R package is to follow the example on the Sonar Wiki and generate a HUGE JSON file from the results (NOTE: this will be over 1TB when some of our newer scans start to be posted). Then, use either jqr or jq on the command-line to extract fields you want/need and then process the resultant JSON in R.

hrbrmstr/r7snr documentation built on May 17, 2019, 5:12 p.m.