Category Archives: Cybersecurity

Acoustic: Solving a CyberDefenders PCAP SIP/RTP Challenge with R, Zeek, tshark (& friends)

2021-07-25 – 09:40
Posted in Cybersecurity, Data Analysis, data driven security, data wrangling, Information Security, pcap, R
Comments (3)

Hot on the heels of the previous CyberDefenders Challenge Solution comes this noisy installment which solves their Acoustic challenge.

You can find the source Rmd on GitHub, but I’m also testing the limits of WP’s markdown rendering and putting it in-stream as well.

No longer book expository this time since much of the setup/explanatory bits from it apply here as well).

Acoustic

Convert the PCAP
Examine and Process log.txt
Process Zeek Logs
Process Packet Summary
What is the transport protocol being used?
The attacker used a bunch of scanning tools that belong to the same suite. Provide the name of the suite.
“What is the User-Agent of the victim system?”
Which tool was only used against the following extensions: 100, 101, 102, 103, and 111?
Which extension on the honeypot does NOT require authentication?
How many extensions were scanned in total?
There is a trace for a real SIP client. What is the corresponding user-agent? (two words, once space in between)
Multiple real-world phone numbers were dialed. Provide the first 11 digits of the number dialed from extension 101?
What are the default credentials used in the attempted basic authentication? (format is username:password)
Which codec does the RTP stream use? (3 words, 2 spaces in between)
How long is the sampling time (in milliseconds)?
What was the password for the account with username 555?
Which RTP packet header field can be used to reorder out of sync
RTP packets in the correct sequence?
The trace includes a secret hidden message. Can you hear
it?

This challenge takes us “into the world of voice communications on the internet. VoIP is becoming the de-facto standard for voice communication. As this technology becomes more common, malicious parties have more opportunities and stronger motives to control these systems to conduct nefarious activities. This challenge was designed to examine and explore some of the attributes of the SIP and RTP protocols.”

We have two files to work with:

log.txt which was generated from an unadvertised, passive honeypot located on the internet such that any traffic destined to it must be nefarious. Unknown parties scanned the honeypot with a range of tools, and this activity is represented in the log file.
- The IP address of the honeypot has been changed to “honey.pot.IP.removed”. In terms of geolocation, pick your favorite city.
- The MD5 hash in the authorization digest is replaced with “MD5_hash_removedXXXXXXXXXXXXXXXX”
- Some octets of external IP addresses have been replaced with an “X”
- Several trailing digits of phone numbers have been replaced with an “X”
- Assume the timestamps in the log files are UTC.
Voip-trace.pcap was created by honeynet members for this forensic challenge to allow participants to employ network analysis skills in the VOIP context.

There are 14 questions to answer.

If you are not familiar with SIP and/or RTP you should do a bit of research first. A good place to start is RTC 3261 (for SIP) and RFC 3550 (for RTC). Some questions may be able to be answered just by knowing the details of these protocols.

Convert the PCAP

library(stringi)
library(tidyverse)

We’ll pre-generate Zeek logs. The -C tells Zeek to not bother with checksums, -r tells it to read from a file and the LogAscii::use_json=T means we want JSON output vs the default delimited files. JSON gives us data types (the headers in the delimited files do as well, but we’d have to write something to read those types then deal with it vs get this for free out of the box with JSON).

system("ZEEK_LOG_SUFFIX=json /opt/zeek/bin/zeek -C -r src/Voip-trace.pcap LogAscii::use_json=T HTTP::default_capture_password=T")

We process the PCAP twice with tshark. Once to get the handy (and small) packet summary table, then dump the whole thing to JSON. We may need to run tshark again down the road a bit.

system("tshark -T tabs -r src/Voip-trace.pcap > voip-packets.tsv")
system("tshark -T json -r src/Voip-trace.pcap > voip-trace")

Examine and Process `log.txt`

We aren’t told what format log.txt is in, so let’s take a look:

cd_sip_log <- stri_read_lines("src/log.txt")

cat(head(cd_sip_log, 25), sep="\n")
## Source: 210.184.X.Y:1083
## Datetime: 2010-05-02 01:43:05.606584
## 
## Message:
## 
## OPTIONS sip:100@honey.pot.IP.removed SIP/2.0
## Via: SIP/2.0/UDP 127.0.0.1:5061;branch=z9hG4bK-2159139916;rport
## Content-Length: 0
## From: "sipvicious"<sip:100@1.1.1.1>; tag=X_removed
## Accept: application/sdp
## User-Agent: friendly-scanner
## To: "sipvicious"<sip:100@1.1.1.1>
## Contact: sip:100@127.0.0.1:5061
## CSeq: 1 OPTIONS
## Call-ID: 845752980453913316694142
## Max-Forwards: 70
## 
## 
## 
## 
## -------------------------
## Source: 210.184.X.Y:4956
## Datetime: 2010-05-02 01:43:12.488811
## 
## Message:

These look a bit like HTTP server responses, but we know we’re working in SIP land and if you perused the RFC you’d have noticed that SIP is an HTTP-like ASCII protocol. While some HTTP response parsers might work on these records, it’s pretty straightforward to whip up a bespoke pseudo-parser.

Let’s see how many records there are by counting the number of “Message:” lines (we’re doing this, primarily, to see if we should use the {furrr} package to speed up processing):

cd_sip_log[stri_detect_fixed(cd_sip_log, "Message:")] %>%
  table()
## .
## Message: 
##     4266

There are many, so we’ll avoid parallel processing the data and just use a single thread.

One way to tackle the parsing is to look for the stop and start of each record, extract fields (these have similar formats to HTTP headers), and perhaps have to extract content as well. We know this because there are “Content-Length:” fields. According to the RFC they are supposed to exist for every message. Let’s first see if any “Content-Length:” header records are greater than 0. We’ll do this with a little help from the ripgrep utility as it provides a way to see context before and/or after matched patterns:

cat(system('rg --after-context=10 "^Content-Length: [^0]" src/log.txt', intern=TRUE), sep="\n")
## Content-Length: 330
## 
## v=0
## o=Zoiper_user 0 0 IN IP4 89.42.194.X
## s=Zoiper_session
## c=IN IP4 89.42.194.X
## t=0 0
## m=audio 52999 RTP/AVP 3 0 8 110 98 101
## a=rtpmap:3 GSM/8000
## a=rtpmap:0 PCMU/8000
## a=rtpmap:8 PCMA/8000
## --
## Content-Length: 330
## 
## v=0
## o=Zoiper_user 0 0 IN IP4 89.42.194.X
## s=Zoiper_session
## c=IN IP4 89.42.194.X
## t=0 0
## m=audio 52999 RTP/AVP 3 0 8 110 98 101
## a=rtpmap:3 GSM/8000
## a=rtpmap:0 PCMU/8000
## a=rtpmap:8 PCMA/8000
## --
## Content-Length: 330
## 
## v=0
## o=Zoiper_user 0 0 IN IP4 89.42.194.X
## s=Zoiper_session
## c=IN IP4 89.42.194.X
## t=0 0
## m=audio 52999 RTP/AVP 3 0 8 110 98 101
## a=rtpmap:3 GSM/8000
## a=rtpmap:0 PCMU/8000
## a=rtpmap:8 PCMA/8000
## --
## Content-Length: 330
## 
## v=0
## o=Zoiper_user 0 0 IN IP4 89.42.194.X
## s=Zoiper_session
## c=IN IP4 89.42.194.X
## t=0 0
## m=audio 52999 RTP/AVP 3 0 8 110 98 101
## a=rtpmap:3 GSM/8000
## a=rtpmap:0 PCMU/8000
## a=rtpmap:8 PCMA/8000

So,we do need to account for content. It’s still pretty straightforward (explanatory comments inline):

starts <- which(stri_detect_regex(cd_sip_log, "^Source:"))
stops <- which(stri_detect_regex(cd_sip_log, "^----------"))

map2_dfr(starts, stops, ~{

  raw_rec <- stri_trim_both(cd_sip_log[.x:.y]) # target the record from the log
  raw_rec <- raw_rec[raw_rec != "-------------------------"] # remove separator

  msg_idx <- which(stri_detect_regex(raw_rec, "^Message:")) # find where "Message:" line is
  source_idx <- which(stri_detect_regex(raw_rec, "^Source: ")) # find where "Source:" line is
  datetime_idx <- which(stri_detect_regex(raw_rec, "^Datetime: ")) # find where "Datetime:" line is
  contents_idx <- which(stri_detect_regex(raw_rec[(msg_idx+2):length(raw_rec)], "^$"))[1] + 2 # get position of the "data"

  source <- stri_match_first_regex(raw_rec[source_idx], "^Source: (.*)$")[,2] # extract source
  datetime <- stri_match_first_regex(raw_rec[datetime_idx], "^Datetime: (.*)$")[,2] # extract datetime
  request <- raw_rec[msg_idx+2] # extract request line

  # build a matrix out of the remaining headers. header key will be in column 2, value will be in column 3
  tmp <- stri_match_first_regex(raw_rec[(msg_idx+3):contents_idx], "^([^:]+):[[:space:]]+(.*)$")
  tmp[,2] <- stri_trans_tolower(tmp[,2]) # lowercase the header key
  tmp[,2] <- stri_replace_all_fixed(tmp[,2], "-", "_") # turn dashes to underscores so we can more easily use the keys as column names

  contents <- raw_rec[(contents_idx+1):length(raw_rec)]
  contents <- paste0(contents[contents != ""], collapse = "\n")

  as.list(tmp[,3]) %>% # turn the header values into a list
    set_names(tmp[,2]) %>% # make their names the tranformed keys
    append(c(
      source = source, # add source to the list (etc)
      datetime = datetime,
      request = request,
      contents = contents
    ))

}) -> sip_log_parsed

Let’s see what we have:

sip_log_parsed
## # A tibble: 4,266 x 18
##    via     content_length from    accept  user_agent to     contact cseq  source
##    <chr>   <chr>          <chr>   <chr>   <chr>      <chr>  <chr>   <chr> <chr> 
##  1 SIP/2.… 0              "\"sip… applic… friendly-… "\"si… sip:10… 1 OP… 210.1…
##  2 SIP/2.… 0              "\"342… applic… friendly-… "\"34… sip:34… 1 RE… 210.1…
##  3 SIP/2.… 0              "\"172… applic… friendly-… "\"17… sip:17… 1 RE… 210.1…
##  4 SIP/2.… 0              "\"adm… applic… friendly-… "\"ad… sip:ad… 1 RE… 210.1…
##  5 SIP/2.… 0              "\"inf… applic… friendly-… "\"in… sip:in… 1 RE… 210.1…
##  6 SIP/2.… 0              "\"tes… applic… friendly-… "\"te… sip:te… 1 RE… 210.1…
##  7 SIP/2.… 0              "\"pos… applic… friendly-… "\"po… sip:po… 1 RE… 210.1…
##  8 SIP/2.… 0              "\"sal… applic… friendly-… "\"sa… sip:sa… 1 RE… 210.1…
##  9 SIP/2.… 0              "\"ser… applic… friendly-… "\"se… sip:se… 1 RE… 210.1…
## 10 SIP/2.… 0              "\"sup… applic… friendly-… "\"su… sip:su… 1 RE… 210.1…
## # … with 4,256 more rows, and 9 more variables: datetime <chr>, request <chr>,
## #   contents <chr>, call_id <chr>, max_forwards <chr>, expires <chr>,
## #   allow <chr>, authorization <chr>, content_type <chr>

glimpse(sip_log_parsed)
## Rows: 4,266
## Columns: 18
## $ via            <chr> "SIP/2.0/UDP 127.0.0.1:5061;branch=z9hG4bK-2159139916;r…
## $ content_length <chr> "0", "0", "0", "0", "0", "0", "0", "0", "0", "0", "0", …
## $ from           <chr> "\"sipvicious\"<sip:100@1.1.1.1>; tag=X_removed", "\"34…
## $ accept         <chr> "application/sdp", "application/sdp", "application/sdp"…
## $ user_agent     <chr> "friendly-scanner", "friendly-scanner", "friendly-scann…
## $ to             <chr> "\"sipvicious\"<sip:100@1.1.1.1>", "\"3428948518\"<sip:…
## $ contact        <chr> "sip:100@127.0.0.1:5061", "sip:3428948518@honey.pot.IP.…
## $ cseq           <chr> "1 OPTIONS", "1 REGISTER", "1 REGISTER", "1 REGISTER", …
## $ source         <chr> "210.184.X.Y:1083", "210.184.X.Y:4956", "210.184.X.Y:51…
## $ datetime       <chr> "2010-05-02 01:43:05.606584", "2010-05-02 01:43:12.4888…
## $ request        <chr> "OPTIONS sip:100@honey.pot.IP.removed SIP/2.0", "REGIST…
## $ contents       <chr> "Call-ID: 845752980453913316694142\nMax-Forwards: 70", …
## $ call_id        <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ max_forwards   <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ expires        <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ allow          <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ authorization  <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ content_type   <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…

Looks ?, but IRL there are edge-cases we’d have to deal with.

Process Zeek Logs

Because they’re JSON files, and the names are reasonable, we can do some magic incantations to read them all in and shove them into a list we’ll call zeek:

zeek <- list()

list.files(
  pattern = "json$",
  full.names = TRUE
) %>%
  walk(~{
    append(zeek, list(file(.x) %>% 
      jsonlite::stream_in(verbose = FALSE) %>%
      as_tibble()) %>% 
        set_names(tools::file_path_sans_ext(basename(.x)))
    ) ->> zeek
  })

str(zeek, 1)
## List of 7
##  $ conn         : tibble [97 × 18] (S3: tbl_df/tbl/data.frame)
##  $ dpd          : tibble [1 × 9] (S3: tbl_df/tbl/data.frame)
##  $ files        : tibble [38 × 16] (S3: tbl_df/tbl/data.frame)
##  $ http         : tibble [92 × 24] (S3: tbl_df/tbl/data.frame)
##  $ packet_filter: tibble [1 × 5] (S3: tbl_df/tbl/data.frame)
##  $ sip          : tibble [9 × 23] (S3: tbl_df/tbl/data.frame)
##  $ weird        : tibble [1 × 9] (S3: tbl_df/tbl/data.frame)

walk2(names(zeek), zeek, ~{
  cat("File:", .x, "\n")
  glimpse(.y)
  cat("\n\n")
})
## File: conn 
## Rows: 97
## Columns: 18
## $ ts            <dbl> 1272737631, 1272737581, 1272737669, 1272737669, 12727376…
## $ uid           <chr> "Cb0OAQ1eC0ZhQTEKNl", "C2s0IU2SZFGVlZyH43", "CcEeLRD3cca…
## $ id.orig_h     <chr> "172.25.105.43", "172.25.105.43", "172.25.105.43", "172.…
## $ id.orig_p     <int> 57086, 5060, 57087, 57088, 57089, 57090, 57091, 57093, 5…
## $ id.resp_h     <chr> "172.25.105.40", "172.25.105.40", "172.25.105.40", "172.…
## $ id.resp_p     <int> 80, 5060, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80…
## $ proto         <chr> "tcp", "udp", "tcp", "tcp", "tcp", "tcp", "tcp", "tcp", …
## $ service       <chr> "http", "sip", "http", "http", "http", "http", "http", "…
## $ duration      <dbl> 0.0180180073, 0.0003528595, 0.0245900154, 0.0740420818, …
## $ orig_bytes    <int> 502, 428, 380, 385, 476, 519, 520, 553, 558, 566, 566, 5…
## $ resp_bytes    <int> 720, 518, 231, 12233, 720, 539, 17499, 144, 144, 144, 14…
## $ conn_state    <chr> "SF", "SF", "SF", "SF", "SF", "SF", "SF", "SF", "SF", "S…
## $ missed_bytes  <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
## $ history       <chr> "ShADadfF", "Dd", "ShADadfF", "ShADadfF", "ShADadfF", "S…
## $ orig_pkts     <int> 5, 1, 5, 12, 5, 6, 16, 6, 6, 5, 5, 5, 5, 5, 5, 5, 6, 5, …
## $ orig_ip_bytes <int> 770, 456, 648, 1017, 744, 839, 1360, 873, 878, 834, 834,…
## $ resp_pkts     <int> 5, 1, 5, 12, 5, 5, 16, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, …
## $ resp_ip_bytes <int> 988, 546, 499, 12865, 988, 807, 18339, 412, 412, 412, 41…
## 
## 
## File: dpd 
## Rows: 1
## Columns: 9
## $ ts             <dbl> 1272737798
## $ uid            <chr> "CADvMziC96POynR2e"
## $ id.orig_h      <chr> "172.25.105.3"
## $ id.orig_p      <int> 43204
## $ id.resp_h      <chr> "172.25.105.40"
## $ id.resp_p      <int> 5060
## $ proto          <chr> "udp"
## $ analyzer       <chr> "SIP"
## $ failure_reason <chr> "Binpac exception: binpac exception: string mismatch at…
## 
## 
## File: files 
## Rows: 38
## Columns: 16
## $ ts             <dbl> 1272737631, 1272737669, 1272737676, 1272737688, 1272737…
## $ fuid           <chr> "FRnb7P5EDeZE4Y3z4", "FOT2gC2yLxjfMCuE5f", "FmUCuA3dzcS…
## $ tx_hosts       <list> "172.25.105.40", "172.25.105.40", "172.25.105.40", "17…
## $ rx_hosts       <list> "172.25.105.43", "172.25.105.43", "172.25.105.43", "17…
## $ conn_uids      <list> "Cb0OAQ1eC0ZhQTEKNl", "CFfYtA0DqqrJk4gI5", "CHN4qA4UUH…
## $ source         <chr> "HTTP", "HTTP", "HTTP", "HTTP", "HTTP", "HTTP", "HTTP",…
## $ depth          <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
## $ analyzers      <list> [], [], [], [], [], [], [], [], [], [], [], [], [], []…
## $ mime_type      <chr> "text/html", "text/html", "text/html", "text/html", "te…
## $ duration       <dbl> 0.000000e+00, 8.920908e-03, 0.000000e+00, 0.000000e+00,…
## $ is_orig        <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, TRUE, FALSE, FALSE, …
## $ seen_bytes     <int> 479, 11819, 479, 313, 17076, 55, 50, 30037, 31608, 1803…
## $ total_bytes    <int> 479, NA, 479, 313, NA, 55, 50, NA, NA, NA, 58, 313, 50,…
## $ missing_bytes  <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
## $ overflow_bytes <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
## $ timedout       <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,…
## 
## 
## File: http 
## Rows: 92
## Columns: 24
## $ ts                <dbl> 1272737631, 1272737669, 1272737669, 1272737676, 1272…
## $ uid               <chr> "Cb0OAQ1eC0ZhQTEKNl", "CcEeLRD3cca3j4QGh", "CFfYtA0D…
## $ id.orig_h         <chr> "172.25.105.43", "172.25.105.43", "172.25.105.43", "…
## $ id.orig_p         <int> 57086, 57087, 57088, 57089, 57090, 57091, 57093, 570…
## $ id.resp_h         <chr> "172.25.105.40", "172.25.105.40", "172.25.105.40", "…
## $ id.resp_p         <int> 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, 80, …
## $ trans_depth       <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1…
## $ method            <chr> "GET", "GET", "GET", "GET", "GET", "GET", "GET", "GE…
## $ host              <chr> "172.25.105.40", "172.25.105.40", "172.25.105.40", "…
## $ uri               <chr> "/maint", "/", "/user/", "/maint", "/maint", "/maint…
## $ referrer          <chr> "http://172.25.105.40/user/", NA, NA, "http://172.25…
## $ version           <chr> "1.1", "1.1", "1.1", "1.1", "1.1", "1.1", "1.1", "1.…
## $ user_agent        <chr> "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.9)…
## $ request_body_len  <int> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
## $ response_body_len <int> 479, 0, 11819, 479, 313, 17076, 0, 0, 0, 0, 0, 0, 0,…
## $ status_code       <int> 401, 302, 200, 401, 301, 200, 304, 304, 304, 304, 30…
## $ status_msg        <chr> "Authorization Required", "Found", "OK", "Authorizat…
## $ tags              <list> [], [], [], [], [], [], [], [], [], [], [], [], [],…
## $ resp_fuids        <list> "FRnb7P5EDeZE4Y3z4", <NULL>, "FOT2gC2yLxjfMCuE5f", …
## $ resp_mime_types   <list> "text/html", <NULL>, "text/html", "text/html", "tex…
## $ username          <chr> NA, NA, NA, NA, "maint", "maint", "maint", "maint", …
## $ password          <chr> NA, NA, NA, NA, "password", "password", "password", …
## $ orig_fuids        <list> <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NU…
## $ orig_mime_types   <list> <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NULL>, <NU…
## 
## 
## File: packet_filter 
## Rows: 1
## Columns: 5
## $ ts      <dbl> 1627151196
## $ node    <chr> "zeek"
## $ filter  <chr> "ip or not ip"
## $ init    <lgl> TRUE
## $ success <lgl> TRUE
## 
## 
## File: sip 
## Rows: 9
## Columns: 23
## $ ts                <dbl> 1272737581, 1272737768, 1272737768, 1272737768, 1272…
## $ uid               <chr> "C2s0IU2SZFGVlZyH43", "CADvMziC96POynR2e", "CADvMziC…
## $ id.orig_h         <chr> "172.25.105.43", "172.25.105.3", "172.25.105.3", "17…
## $ id.orig_p         <int> 5060, 43204, 43204, 43204, 43204, 43204, 43204, 4320…
## $ id.resp_h         <chr> "172.25.105.40", "172.25.105.40", "172.25.105.40", "…
## $ id.resp_p         <int> 5060, 5060, 5060, 5060, 5060, 5060, 5060, 5060, 5060
## $ trans_depth       <int> 0, 0, 0, 0, 0, 0, 0, 0, 0
## $ method            <chr> "OPTIONS", "REGISTER", "REGISTER", "SUBSCRIBE", "SUB…
## $ uri               <chr> "sip:100@172.25.105.40", "sip:172.25.105.40", "sip:1…
## $ request_from      <chr> "\"sipvicious\"<sip:100@1.1.1.1>", "<sip:555@172.25.…
## $ request_to        <chr> "\"sipvicious\"<sip:100@1.1.1.1>", "<sip:555@172.25.…
## $ response_from     <chr> "\"sipvicious\"<sip:100@1.1.1.1>", "<sip:555@172.25.…
## $ response_to       <chr> "\"sipvicious\"<sip:100@1.1.1.1>;tag=as18cdb0c9", "<…
## $ call_id           <chr> "61127078793469957194131", "MzEwMmYyYWRiYTUxYTBhODY3…
## $ seq               <chr> "1 OPTIONS", "1 REGISTER", "2 REGISTER", "1 SUBSCRIB…
## $ request_path      <list> "SIP/2.0/UDP 127.0.1.1:5060", "SIP/2.0/UDP 172.25.10…
## $ response_path     <list> "SIP/2.0/UDP 127.0.1.1:5060", "SIP/2.0/UDP 172.25.10…
## $ user_agent        <chr> "UNfriendly-scanner - for demo purposes", "X-Lite B…
## $ status_code       <int> 200, 401, 200, 401, 404, 401, 100, 200, NA
## $ status_msg        <chr> "OK", "Unauthorized", "OK", "Unauthorized", "Not fo…
## $ request_body_len  <int> 0, 0, 0, 0, 0, 264, 264, 264, 0
## $ response_body_len <int> 0, 0, 0, 0, 0, 0, 0, 302, NA
## $ content_type      <chr> NA, NA, NA, NA, NA, NA, NA, "application/sdp", NA
## 
## 
## File: weird 
## Rows: 1
## Columns: 9
## $ ts        <dbl> 1272737805
## $ id.orig_h <chr> "172.25.105.3"
## $ id.orig_p <int> 0
## $ id.resp_h <chr> "172.25.105.40"
## $ id.resp_p <int> 0
## $ name      <chr> "truncated_IPv6"
## $ notice    <lgl> FALSE
## $ peer      <chr> "zeek"
## $ source    <chr> "IP"

Process Packet Summary

We won’t process the big JSON file tshark generated for us util we really have to, but we can read in the packet summary table now:

packet_cols <- c("packet_num", "ts", "src", "discard", "dst", "proto", "length", "info")

read_tsv(
  file = "voip-packets.tsv",
  col_names = packet_cols,
  col_types = "ddccccdc"
) %>%
  select(-discard) -> packets

packets
## # A tibble: 4,447 x 7
##    packet_num       ts src      dst     proto length info                       
##         <dbl>    <dbl> <chr>    <chr>   <chr>  <dbl> <chr>                      
##  1          1  0       172.25.… 172.25… SIP      470 Request: OPTIONS sip:100@1…
##  2          2  3.53e-4 172.25.… 172.25… SIP      560 Status: 200 OK |           
##  3          3  5.03e+1 172.25.… 172.25… TCP       74 57086 → 80 [SYN] Seq=0 Win…
##  4          4  5.03e+1 172.25.… 172.25… TCP       74 80 → 57086 [SYN, ACK] Seq=…
##  5          5  5.03e+1 172.25.… 172.25… TCP       66 57086 → 80 [ACK] Seq=1 Ack…
##  6          6  5.03e+1 172.25.… 172.25… HTTP     568 GET /maint HTTP/1.1        
##  7          7  5.03e+1 172.25.… 172.25… TCP       66 80 → 57086 [ACK] Seq=1 Ack…
##  8          8  5.03e+1 172.25.… 172.25… HTTP     786 HTTP/1.1 401 Authorization…
##  9          9  5.03e+1 172.25.… 172.25… TCP       66 80 → 57086 [FIN, ACK] Seq=…
## 10         10  5.03e+1 172.25.… 172.25… TCP       66 57086 → 80 [ACK] Seq=503 A…
## # … with 4,437 more rows

glimpse(packets)
## Rows: 4,447
## Columns: 7
## $ packet_num <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, …
## $ ts         <dbl> 0.000000, 0.000353, 50.317176, 50.317365, 50.320071, 50.329…
## $ src        <chr> "172.25.105.43", "172.25.105.40", "172.25.105.43", "172.25.…
## $ dst        <chr> "172.25.105.40", "172.25.105.43", "172.25.105.40", "172.25.…
## $ proto      <chr> "SIP", "SIP", "TCP", "TCP", "TCP", "HTTP", "TCP", "HTTP", "…
## $ length     <dbl> 470, 560, 74, 74, 66, 568, 66, 786, 66, 66, 66, 66, 74, 74,…
## $ info       <chr> "Request: OPTIONS sip:100@172.25.105.40 |", "Status: 200 OK…

What is the transport protocol being used?

SIP can use TCP or UDP and which transport it uses will be specified in the Via: header. Let’s take a look:

head(sip_log_parsed$via)
## [1] "SIP/2.0/UDP 127.0.0.1:5061;branch=z9hG4bK-2159139916;rport"
## [2] "SIP/2.0/UDP 127.0.0.1:5087;branch=z9hG4bK-1189344537;rport"
## [3] "SIP/2.0/UDP 127.0.0.1:5066;branch=z9hG4bK-2119091576;rport"
## [4] "SIP/2.0/UDP 127.0.0.1:5087;branch=z9hG4bK-3226446220;rport"
## [5] "SIP/2.0/UDP 127.0.0.1:5087;branch=z9hG4bK-1330901245;rport"
## [6] "SIP/2.0/UDP 127.0.0.1:5087;branch=z9hG4bK-945386205;rport"

Are they all UDP? We can find out by performing some light processing
on the via column:

sip_log_parsed %>% 
  select(via) %>% 
  mutate(
    transport = stri_match_first_regex(via, "^([^[:space:]]+)")[,2]
  ) %>% 
  count(transport, sort=TRUE)
## # A tibble: 1 x 2
##   transport       n
##   <chr>       <int>
## 1 SIP/2.0/UDP  4266

Looks like they’re all UDP. Question 1: ✅

The attacker used a bunch of scanning tools that belong to the same suite. Provide the name of the suite.

Don’t you, now, wish you had listen to your parents when they were telling you about the facts of SIP life when you were a wee pup?

We’ll stick with the SIP log to answer this one and peek back at the RFC to see that there’s a “User-Agent:” field which contains information about the client originating the request. Most scanners written by defenders identify themselves in User-Agent fields when those fields are available in a protocol exchange, and a large percentage of naive malicious folks are too daft to change this value (or leave it default to make you think they’re not behaving badly).

If you are a regular visitor to SIP land, you likely know the common SIP scanning tools. These are a few:

Nmap’s SIP library
Mr.SIP, a “SIP-Based Audit and Attack Tool”
SIPVicious, a “set of security tools that can be used to audit SIP based VoIP systems”
Sippts, a “set of tools to audit SIP based VoIP Systems”

(There are many more.)

Let’s see what user-agent was used in this log extract:

count(sip_log_parsed, user_agent, sort=TRUE)
## # A tibble: 3 x 2
##   user_agent           n
##   <chr>            <int>
## 1 friendly-scanner  4248
## 2 Zoiper rev.6751     14
## 3 <NA>                 4

The overwhelming majority are friendly-scanner. Let’s look at a few of those log entries:

sip_log_parsed %>% 
  filter(
    user_agent == "friendly-scanner"
  ) %>% 
  glimpse()
## Rows: 4,248
## Columns: 18
## $ via            <chr> "SIP/2.0/UDP 127.0.0.1:5061;branch=z9hG4bK-2159139916;r…
## $ content_length <chr> "0", "0", "0", "0", "0", "0", "0", "0", "0", "0", "0", …
## $ from           <chr> "\"sipvicious\"<sip:100@1.1.1.1>; tag=X_removed", "\"34…
## $ accept         <chr> "application/sdp", "application/sdp", "application/sdp"…
## $ user_agent     <chr> "friendly-scanner", "friendly-scanner", "friendly-scann…
## $ to             <chr> "\"sipvicious\"<sip:100@1.1.1.1>", "\"3428948518\"<sip:…
## $ contact        <chr> "sip:100@127.0.0.1:5061", "sip:3428948518@honey.pot.IP.…
## $ cseq           <chr> "1 OPTIONS", "1 REGISTER", "1 REGISTER", "1 REGISTER", …
## $ source         <chr> "210.184.X.Y:1083", "210.184.X.Y:4956", "210.184.X.Y:51…
## $ datetime       <chr> "2010-05-02 01:43:05.606584", "2010-05-02 01:43:12.4888…
## $ request        <chr> "OPTIONS sip:100@honey.pot.IP.removed SIP/2.0", "REGIST…
## $ contents       <chr> "Call-ID: 845752980453913316694142\nMax-Forwards: 70", …
## $ call_id        <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ max_forwards   <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ expires        <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ allow          <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ authorization  <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
## $ content_type   <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…

Those from and to fields have an interesting name in them: “sipviscious”. You’ve seen that before, right at the beginning of this section.

Let’s do a quick check over at the SIPvicious repo just to make sure.

count(sip_log_parsed, user_agent)
## # A tibble: 3 x 2
##   user_agent           n
##   <chr>            <int>
## 1 friendly-scanner  4248
## 2 Zoiper rev.6751     14
## 3 <NA>                 4

“What is the User-Agent of the victim system?”

We only have partial data in the text log so we’ll have to look elsewhere (the PCAP) for this information. The “victim” is whatever was the target of a this SIP-based attack and we can look for SIP messages, user agents, and associated IPs in the PCAP thanks to tshark’s rich SIP filter library:

system("tshark -Q -T fields -e ip.src -e ip.dst -e sip.User-Agent -r src/Voip-trace.pcap 'sip.User-Agent'")

That first exchange is all we really need. We see our rude poker talking to 172.25.105.40 and it responding right after.

Which tool was only used against the following extensions: 100, 101, 102, 103, and 111?

The question is a tad vague and is assuming — since we now know the SIPvicious suite was used — that we also know to provide the name of the Python script in SIPvicious that was used. There are five tools:

svmap: this is a sip scanner. When launched against ranges of ip address space, it will identify any SIP servers which it finds on the way. Also has the option to scan hosts on ranges of ports. Usage: https://github.com/EnableSecurity/sipvicious/wiki/SVMap-Usage
svwar: identifies working extension lines on a PBX. A working extension is one that can be registered. Also tells you if the extension line requires authentication or not. Usage: https://github.com/EnableSecurity/sipvicious/wiki/SVWar-Usage
svcrack: a password cracker making use of digest authentication. It is able to crack passwords on both registrar servers and proxy servers. Current cracking modes are either numeric ranges or words from dictionary files. Usage: https://github.com/EnableSecurity/sipvicious/wiki/SVCrack-Usage
svreport: able to manage sessions created by the rest of the tools and export to pdf, xml, csv and plain text. Usage: https://github.com/EnableSecurity/sipvicious/wiki/SVReport-Usage
svcrash: responds to svwar and svcrack SIP messages with a message that causes old versions to crash. Usage: https://github.com/EnableSecurity/sipvicious/wiki/SVCrash-FAQ

The svcrash tool is something defenders can use to help curtail scanner activity. We can cross that off the list. The svreport tool is for working with data generated by svmap, svwar and/or svcrack. One more crossed off. We also know that the attacker scanned the SIP network looking for nodes, which means svmap and svwar are likely not exclusive tool to the target extensions. (We technically have enough information right now to answer the question especially if you look carefully at the answer box on the site but that’s cheating).

The SIP request line and header field like “To:” destination information in the form of a SIP URI. Since we only care about the extension component of the URI for this question, we can use a regular expression to isolate them.

Back to the SIP log to see if we can find the identified extensions. We’ll also process the “From:” header just in case we need it.

sip_log_parsed %>% 
  mutate_at(
    vars(request, from, to),
    ~stri_match_first_regex(.x, "sip:([^@]+)@")[,2]
  ) %>% 
  select(request, from, to)
## # A tibble: 4,266 x 3
##    request    from       to        
##    <chr>      <chr>      <chr>     
##  1 100        100        100       
##  2 3428948518 3428948518 3428948518
##  3 1729240413 1729240413 1729240413
##  4 admin      admin      admin     
##  5 info       info       info      
##  6 test       test       test      
##  7 postmaster postmaster postmaster
##  8 sales      sales      sales     
##  9 service    service    service   
## 10 support    support    support   
## # … with 4,256 more rows

That worked! We can now see what friendly-scanner attempted to authenticate only to our targets:

sip_log_parsed %>%
  mutate_at(
    vars(request, from, to),
    ~stri_match_first_regex(.x, "sip:([^@]+)@")[,2]
  ) %>% 
  filter(
    user_agent == "friendly-scanner",
    stri_detect_fixed(contents, "Authorization")
  ) %>% 
  distinct(to)
## # A tibble: 4 x 1
##   to   
##   <chr>
## 1 102  
## 2 103  
## 3 101  
## 4 111

While we’re missing 100 that’s likely due to it not requiring authentication (svcrack will REGISTER first to determine if a target requires authentication and not send cracking requests if it doesn’t).

Which extension on the honeypot does NOT require authentication?

We know this due to what we found in the previous question. Extension 100 does not require authentication.

How many extensions were scanned in total?

We just need to count the distinct to’s where the user agent is the scanner:

sip_log_parsed %>% 
  mutate_at(
    vars(request, from, to),
    ~stri_match_first_regex(.x, "sip:([^@]+)@")[,2]
  ) %>% 
  filter(
    user_agent == "friendly-scanner"
  ) %>% 
  distinct(to)
## # A tibble: 2,652 x 1
##    to        
##    <chr>     
##  1 100       
##  2 3428948518
##  3 1729240413
##  4 admin     
##  5 info      
##  6 test      
##  7 postmaster
##  8 sales     
##  9 service   
## 10 support   
## # … with 2,642 more rows

There is a trace for a real SIP client. What is the corresponding user-agent? (two words, once space in between)

We only need to look for user agent’s that aren’t our scanner:

sip_log_parsed %>% 
  filter(
    user_agent != "friendly-scanner"
  ) %>% 
  count(user_agent)
## # A tibble: 1 x 2
##   user_agent          n
##   <chr>           <int>
## 1 Zoiper rev.6751    14

Multiple real-world phone numbers were dialed. Provide the first 11 digits of the number dialed from extension 101?

Calls are “INVITE” requests

sip_log_parsed %>% 
  mutate_at(
    vars(from, to),
    ~stri_match_first_regex(.x, "sip:([^@]+)@")[,2]
  ) %>% 
  filter(
    from == 101,
    stri_detect_regex(cseq, "INVITE")
  ) %>% 
  select(to) 
## # A tibble: 3 x 1
##   to              
##   <chr>           
## 1 900114382089XXXX
## 2 00112322228XXXX 
## 3 00112524021XXXX

The challenge answer box provides hint to what number they want. I’m not sure but I suspect it may be randomized, so you’ll have to match the pattern they expect with the correct digits above.

What are the default credentials used in the attempted basic authentication? (format is username:password)

This question wants us to look at the HTTP requests that require authentication. We can get he credentials info from the zeek$http log:

zeek$http %>% 
  distinct(username, password)
## # A tibble: 2 x 2
##   username password
##   <chr>    <chr>   
## 1 <NA>     <NA>    
## 2 maint    password

Which codec does the RTP stream use? (3 words, 2 spaces in between)

“Codec” refers to the algorithm used to encode/decode an audio or video stream. The RTP RFC uses the term “payload type” to refer to this during exchanges and even has a link to RFC 3551 which provides further information on these encodings.

The summary packet table that tshark generates helpfully provides summary info for RTP packets and part of that info is PT=… which indicates the payload type.

packets %>% 
  filter(proto == "RTP") %>% 
  select(info)
## # A tibble: 2,988 x 1
##    info                                                       
##    <chr>                                                      
##  1 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6402, Time=126160
##  2 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6403, Time=126320
##  3 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6404, Time=126480
##  4 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6405, Time=126640
##  5 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6406, Time=126800
##  6 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6407, Time=126960
##  7 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6408, Time=127120
##  8 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6409, Time=127280
##  9 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6410, Time=127440
## 10 PT=ITU-T G.711 PCMU, SSRC=0xA254E017, Seq=6411, Time=127600
## # … with 2,978 more rows

How long is the sampling time (in milliseconds)?

1 Hz = 1,000 ms
1 ms = 1,000 Hz

(1/8000) * 1000

What was the password for the account with username 555?

We don’t really need to use external programs for this but it will sure go quite a bit faster if we do. While the original reference page for sipdump and sipcrack is defunct, you can visit that link to go to the Wayback machine’s capture of it. It will help if you have a linux system handy (so Docker to the rescue for macOS and Windows folks) since the following answer details are running on Ubunbu.

This question is taking advantage of the fact that the default authentication method for SIP is extremely weak. The process uses an MD5 challenge/response, and if an attacker can capture call traffic it is possible to brute force the password offline (which is what we’ll use sipcrack for).

You can install them via sudo apt install sipcrack.

We’ll first generate a dump of the authentication attempts with sipdump:

system("sipdump -p src/Voip-trace.pcap sip.dump", intern=TRUE)
##  [1] ""                                                               
##  [2] "SIPdump 0.2 "                                                   
##  [3] "---------------------------------------"                        
##  [4] ""                                                               
##  [5] "* Using pcap file 'src/Voip-trace.pcap' for sniffing"           
##  [6] "* Starting to sniff with packet filter 'tcp or udp'"            
##  [7] ""                                                               
##  [8] "* Dumped login from 172.25.105.40 -> 172.25.105.3 (User: '555')"
##  [9] "* Dumped login from 172.25.105.40 -> 172.25.105.3 (User: '555')"
## [10] "* Dumped login from 172.25.105.40 -> 172.25.105.3 (User: '555')"
## [11] ""                                                               
## [12] "* Exiting, sniffed 3 logins"

cat(readLines("sip.dump"), sep="\n")
## 172.25.105.3"172.25.105.40"555"asterisk"REGISTER"sip:172.25.105.40"4787f7ce""""MD5"1ac95ce17e1f0230751cf1fd3d278320
## 172.25.105.3"172.25.105.40"555"asterisk"INVITE"sip:1000@172.25.105.40"70fbfdae""""MD5"aa533f6efa2b2abac675c1ee6cbde327
## 172.25.105.3"172.25.105.40"555"asterisk"BYE"sip:1000@172.25.105.40"70fbfdae""""MD5"0b306e9db1f819dd824acf3227b60e07

It saves the IPs, caller, authentication realm, method, nonce and hash which will all be fed into the sipcrack.

We know from the placeholder answer text that the “password” is 4 characters, and this is the land of telephony, so we can make an assumption that it is really 4 digits. sipcrack needs a file of passwords to try, so We’ll let R make a randomized file of 4 digit pins for us:

cat(sprintf("%04d", sample(0:9999)), file = "4-digits", sep="\n")

We only have authenticaton packets for 555 so we can automate what would normally be an interactive process:

cat(system('echo "1" | sipcrack -w 4-digits sip.dump', intern=TRUE), sep="\n")
## 
## SIPcrack 0.2 
## ----------------------------------------
## 
## * Found Accounts:
## 
## Num  Server      Client      User    Hash|Password
## 
## 1    172.25.105.3    172.25.105.40   555 1ac95ce17e1f0230751cf1fd3d278320
## 2    172.25.105.3    172.25.105.40   555 aa533f6efa2b2abac675c1ee6cbde327
## 3    172.25.105.3    172.25.105.40   555 0b306e9db1f819dd824acf3227b60e07
## 
## * Select which entry to crack (1 - 3): 
## * Generating static MD5 hash... c3e0f1664fde9fbc75a7cbd341877875
## * Loaded wordlist: '4-digits'
## * Starting bruteforce against user '555' (MD5: '1ac95ce17e1f0230751cf1fd3d278320')
## * Tried 8904 passwords in 0 seconds
## 
## * Found password: '1234'
## * Updating dump file 'sip.dump'... done

Which RTP packet header field can be used to reorder out of sync RTP packets in the correct sequence?

Just reading involved here: 5.1 RTP Fixed Header Fields.

The trace includes a secret hidden message. Can you hear it?

We could command line this one but honestly Wireshark has a pretty keen audio player. Fire it up, open up the PCAP, go to the “Telephony” menu, pick SIP and play the streams.

Packet Maze: Solving a CyberDefenders PCAP Puzzle with R, Zeek, and tshark

2021-07-20 – 15:18
Posted in Cybersecurity, Data Analysis, data driven security, data science, data wrangling, DNS, Information Security, pcap, R
Comments (7)

It was a rainy weekend in southern Maine and I really didn’t feel like doing chores, so I was skimming through RSS feeds and noticed a link to a PacketMaze challenge in the latest This Week In 4n6.

Since it’s also been a while since I’ve done any serious content delivery (on the personal side, anyway), I thought it’d be fun to solve the challenge with some tools I like — namely Zeek, tshark, and R (links to those in the e-book I’m linking to below), craft some real expository around each solution, and bundle it all up into an e-book and lighter-weight GitHub repo.

There are 11 “quests” in the challenge, requiring sifting through a packet capture (PCAP) and looking for various odds and ends (some are very windy maze passages). The challenge ranges from extracting images and image metadata from FTP sessions to pulling out precise elements in TLS sessions, to dealing with IPv6.

This is far from an expert challenge, and anyone can likely work through it with a little bit of elbow grease.

As it says on the tin, not all data is ‘big’ nor do all data-driven cybersecurity projects require advanced modeling capabilities. Sometimes you just need to dissect some network packet capture (PCAP) data and don’t want to click through a GUI to get the job done. This short book works through the questions in CyberDefenders Lab #68 to show how you can get the Zeek open source network security tool, tshark command-line PCAP analysis Swiss army knife, and R (via RStudio) working together.

FIN

If you find the resource helpful or have other feedback, drop a note on Twitter (@hrbrmstr), in a comment here, or as a GitHub issue.

Feedly Mini Extension Removed From Chrome Store Due To “Malware”

2021-05-09 – 07:57
Posted in Cybersecurity, Feedly, Malware
Leave a Comment

On or about Friday evening (May 7, 2021) Edge notified me that the Feedly Mini extension (one of the only extensions I use as extensions are dangerous things) was remove from the store due to “malware”.

Feedly is used by many newshounds, and with 2021 being a very bad year when it comes to supply-chain attacks, seeing a notice about malware in a very popular Chrome extension is more than a little distressing.

I’m posting this blog to get the word “malware” associated with “Feedly” so they are compelled to make some sort of statement. I’ll update it with more information as it is provided.

Help Your Mac Stand Between The Darkness And The Light with GreyWatch

2021-03-30 – 06:57
Posted in Cybersecurity, macOS, Objective-C, Swift
Leave a Comment

Greynoise helps security teams focus on potential threats by reducing the noise from logs, alerts, and SIEMs. They constantly watch for badly behaving internet hosts, keep track of the benign ones, and use this research to classify IP addresses. Teams can use these classifications to only focus on things that (potentially) matter.

They also have a generous (10K calls/day), free community API which does not require credentialed access and returns a subset of information that the full API does. This is handy for folks who can’t afford the service or who only need to occasionally poke at IP addresses.

Andrew, GN’s CEO, tweeted out a super-hacky shell one-liner, the other day, that grabs the external IPs of all the ESTABLISHED IPv4 TCP connections and runs them through the community API via curl. Even though I made it a bit less-hacky:

sudo netstat -anp TCP \
  | rg ESTAB \
  | rg "(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)" -o \
  | rg -v "(^127\.)|(^10\.)|(^172\.1[6-9]\.)|(^172\.2[0-9]\.)|(^172\.3[0-1]\.)|(^192\.168\.)" \
  | rg -v "$(dig +short viz.greynoise.io @9.9.9.9 | rg '^\d' | tr '\n' '|' | sed -e 's/.$//g')" \
  | sort -u \
  | while read IP; do echo $(curl --silent https://api.greynoise.io/v3/community/$IP); done |
  Rscript -e 'tibble::as_tibble(jsonlite::stream_in(file("stdin"), verbose=FALSE))'

its still a “run-on-demand” process that you could put in a script and launchd, but then you’d still have to keep a terminal up or remember to watch some file. Plus, it relies on full executables.

I decided to make things a bit easier for folks on macOS Big Sur by cranking out a small SwiftUI app I’ve dubbed GreyWatch:

Each list entry show an IP address your Mac previously connected to (since app launch) or currently has established TCP connections to. The three indicator dots show (in order) whether Greynoise has detected scanning behavior from the IP address within the last 30 days, whether it has a “Rule It OuT” (RIOT) classification, and what — if any — classification the IP address has. The app only shows an IP address once even it you continue to connect to it and it puts new connections on top.

If an IP address has a classification, double-clicking it will open your default browser to the Greynoise visualizer, otherwise said double-click will take you to the IPInfo entry for the IP address.

Needless to say, if your Mac is talking to a host Greynoise has classified as horribad, your other 99 problems no longer take precedence. I’ll likely add a notification action if that condition occurrs.

There’s an “Export…” item in the file menu that lets you save a copy of the current IP list (with metadata) to an ndlines formatted JSON file.

The app does not shell out to dig or netstat and has a light memory and energy footprint.

There are pre-built, notarized binaries in the releases section, and I’ll gradually be adding features (submit yours via new issues!). You can also submit bug reports or other questions via GH issues as well.

Many thanks to Andrew and team for their generous free tier, which enables semi-useful community hacks like this one!

Brimming With Possibilities: Query zqd & Mine Logs with zq from R

Brim Security maintains a free, Electron-based desktop GUI for exploration of PCAPs and select cybersecurity logs:

along with a broad ecosystem of tools which can be used independently of the GUI.

The standalone or embedded zqd server, as well as the zq command line utility let analysts run ZQL (a domain-specific query language) queries on cybersecurity data sources.

The Brim team maintains a Python module that is capable of working with the zqd HTTP API and my nascent {brimr}^{gitea|gh|gl|bb} R package provides a similar API structure to perform similar operations in R, along with a wrapper for the zq commmand line tool.

PCAPs! In! Spaaaaacce[s]!

Brim Desktop organizes input sources into something called “spaces”. We can check for available spaces with brim_spaces():

library(brimr)
library(tibble)

brim_spaces()
##                               id                                                            name
## 1 sp_1p6pwLgtsESYBTHU9PL9fcl2iBn 2021-02-17-Trickbot-gtag-rob13-infection-in-AD-environment.pcap
##                                                                                              data_path storage_kind
## 1 file:///Users/demo/Library/Application%20Support/Brim/data/spaces/sp_1p6pwLgtsESYBTHU9PL9fcl2iBn    filestore

This single space availble is a sample capture of Trickbot

Let’s profile the network connections in this capture:

# ZQL query to fetch Zeek connection data
zql1 <- '_path=conn | count() by id.orig_h, id.resp_h, id.resp_p | sort id.orig_h, id.resp_h, id.resp_p'

space <- "2021-02-17-Trickbot-gtag-rob13-infection-in-AD-environment.pcap"

r1 <- brim_search(space, zql1)

r1
## ZQL query took 0.0000 seconds; 384 records matched; 1,082 records read; 238,052 bytes read

(r1 <- as_tibble(tidy_brim(r1)))
## # A tibble: 74 x 4
##    orig_h      resp_h       resp_p count
##    <chr>       <chr>        <chr>  <int>
##  1 10.2.17.2   10.2.17.101  49787      1
##  2 10.2.17.101 3.222.126.94 80         1
##  3 10.2.17.101 10.2.17.1    445        1
##  4 10.2.17.101 10.2.17.2    53        97
##  5 10.2.17.101 10.2.17.2    88        27
##  6 10.2.17.101 10.2.17.2    123        5
##  7 10.2.17.101 10.2.17.2    135        8
##  8 10.2.17.101 10.2.17.2    137        2
##  9 10.2.17.101 10.2.17.2    138        2
## 10 10.2.17.101 10.2.17.2    389       37
## # … with 64 more rows

Brim auto-processed the PCAP into Zeek log format and _path=conn in query string indicates that’s where we’re going to perform further data operations (the queries are structured a bit like jq filters). We then ask Brim/zqd to summarize and sort source IP, destination IP, and port counts. {brimr} sends this query over to the server. The raw response is a custom data structure that we can turn into a tidy data frame via tidy_brim().

We can do something similar with the Suricata data that Brim also auto-processes for us:

# Z query to fetch Suricata alerts including the count of alerts per source:destination 
zql2 <- "event_type=alert | count() by src_ip, dest_ip, dest_port, alert.severity, alert.signature | sort src_ip, dest_ip, dest_port, alert.severity, alert.signature"

r2 <- brim_search(space, zql2)

r2
## ZQL query took 0.0000 seconds; 47 records matched; 870 records read; 238,660 bytes read

(r2 <- (as_tibble(tidy_brim(r2))))
## # A tibble: 35 x 6
##    src_ip     dest_ip    dest_port severity signature                                                              count
##    <chr>      <chr>          <int>    <int> <chr>                                                                  <int>
##  1 10.2.17.2  10.2.17.1…     49674        3 SURICATA Applayer Detect protocol only one direction                       1
##  2 10.2.17.2  10.2.17.1…     49680        3 SURICATA Applayer Detect protocol only one direction                       1
##  3 10.2.17.2  10.2.17.1…     49687        3 SURICATA Applayer Detect protocol only one direction                       1
##  4 10.2.17.2  10.2.17.1…     49704        3 SURICATA Applayer Detect protocol only one direction                       1
##  5 10.2.17.2  10.2.17.1…     49709        3 SURICATA Applayer Detect protocol only one direction                       1
##  6 10.2.17.2  10.2.17.1…     49721        3 SURICATA Applayer Detect protocol only one direction                       1
##  7 10.2.17.2  10.2.17.1…     50126        3 SURICATA Applayer Detect protocol only one direction                       1
##  8 10.2.17.1… 3.222.126…        80        2 ET POLICY curl User-Agent Outbound                                         1
##  9 10.2.17.1… 36.95.27.…       443        1 ET HUNTING Suspicious POST with Common Windows Process Names - Possib…     1
## 10 10.2.17.1… 36.95.27.…       443        1 ET MALWARE Win32/Trickbot Data Exfiltration                                1
## # … with 25 more rows

Finally, for this toy example, we’ll also generate a visual overview of these connections:

library(igraph)
library(ggraph)
library(tidyverse)

gdf <- count(r1, orig_h, resp_h, wt=count)

count(gdf, node = resp_h, wt=n, name = "in_degree") %>% 
  full_join(
    count(gdf, node = orig_h, name = "out_degree")
  ) %>% 
  mutate_at(
    vars(in_degree, out_degree),
    replace_na, 1
  ) %>% 
  arrange(in_degree) -> vdf

g <- graph_from_data_frame(gdf, vertices = vdf)

ggraph(g, layout = "linear") +
  geom_node_point(
    aes(size = in_degree), shape = 21
  ) +
  geom_edge_arc(
    width = 0.125, 
    arrow = arrow(
      length = unit(5, "pt"),
      type = "closed"
    )
  )

We can also process log files directly (i.e. without any server) with zq_cmd():

zq_cmd(
  c(
    '"* | cut ts,id.orig_h,id.orig_p"', # note the quotes
    system.file("logs", "conn.log.gz", package = "brimr")
   )
 )
##           id.orig_h id.orig_p                          ts
##   1:  10.164.94.120     39681 2018-03-24T17:15:21.255387Z
##   2:    10.47.25.80     50817 2018-03-24T17:15:21.411148Z
##   3:    10.47.25.80     50817 2018-03-24T17:15:21.926018Z
##   4:    10.47.25.80     50813 2018-03-24T17:15:22.690601Z
##   5:    10.47.25.80     50813 2018-03-24T17:15:23.205187Z
##  ---                                                     
## 988: 10.174.251.215     33003 2018-03-24T17:15:21.429238Z
## 989: 10.174.251.215     33003 2018-03-24T17:15:21.429315Z
## 990: 10.174.251.215     33003 2018-03-24T17:15:21.429479Z
## 991:  10.164.94.120     38265 2018-03-24T17:15:21.427375Z
## 992: 10.174.251.215     33003 2018-03-24T17:15:21.433306Z

FIN

This package is less than 24 hrs old (as of the original blog post date) and there are still a few bits missing, which means y’all have the ability to guide the direction it heads in. So kick the tyres and interact where you’re most comfortable.

A Look at PAN-OS Versions with a Bit of R

2020-07-10 – 11:03
Posted in Cybersecurity, Information Security, R
Comments (2)

The incredibly talented folks over at Bishop Fox were quite generous this week, providing a scanner for figuring out PAN-OS GlobalProtect versions. I’ve been using their decoding technique and date-based fingerprint table to keep an eye on patch status (over at $DAYJOB we help customers, organizations, and national cybersecurity centers get ahead of issues as best as we can).

We have at-scale platforms for scanning the internet and aren’t running the panos-scanner repo code, but I since there is Python code for doing this, I thought it might be fun to show R folks how to do the same thing and show folks how to use the {httr} package to build something similar (we won’t hit all the URLs their script assesses, primarily for brevity).

What Are We Doing Again?

Palo Alto makes many things, most of which are built on their custom linux distribution dubbed PAN-OS. One of the things they build is a VPN product that lets users remotely access internal company resources. It’s had quite the number of pretty horrible problems of late.

Folks contracted to assess security defenses (colloquially dubbed “pen-testers” tho no pens are usually involved) and practitioners within an organization who want to gain an independent view of what their internet perimeter looks like often assemble tools to perform
ad-hoc assessments. Sure, there are commercial tools for performing these assessments ($DAYJOB makes some!), but these open source tools make it possible for folks to learn from each other and for makers of products (like PAN-OS) to do a better job securing their creations.

In this case, the creation is a script that lets the caller figure out what version of PAN-OS is running on a given GlobalProtect box.

To follow along at home, you’ll need access to a PAN-OS system, as I’m not providing an IP address of one for you. It’s really not hard to find one (just google a bit or stand up a trial one from the vendor). Throughout the examples I’ll be using {glue} to replace ip and port in various function calls, so let’s get some setup bits out of the way:

library(httr)
library(tidyverse) # for read_fwf() (et al), pluck(), filter(), and %>%

gg <- glue::glue # no need to bring in the entire namespace just for this

Assuming you have a valid ip and port, let’s try making a request against your PAN-OS GlobalProtect (hereafter using “GP” so save keystrokes) system:

httr::HEAD(
  url = gg("https://{ip}:{port}/global-protect/login.esp")
) -> res
## Error in curl::curl_fetch_memory(url, handle = handle) : 
##  SSL certificate problem: self signed certificate

We’re using a HEAD request as we really don’t need the contents of the remote file (unless you need to verify it truly is a PAN-OS GP server), just the metadata about it. You can use a traditional GET request if you like, though.

We immediately run into a snag since these boxes tend to use a self-signed SSL/TLS certificate which web clients aren’t thrilled about dealing with unless explicitly configured to. We can circumvent this with some configuration options, but you should not use the following incantations haphazardly. SSL/TLS no longer really means what it used to (thanks, Let’s Encrypt!) but you have no guarantees of what’s being delivered to you is legitimate if you hit a plaintext web site or one with an invalid certificate. Enough with the soapbox, let’s make the request:

httr::HEAD(
  url = gg("https://{ip}:{port}/global-protect/login.esp"),
  config = httr::config(
    ssl_verifyhost =FALSE, 
    ssl_verifypeer = FALSE
  )
) -> res

httr::status_code(res)
## [1] 200

In that request, we’ve told the underlying {curl} library calls to not verify the validity of the host or peer certificates associated with the service. Again, don’t do this haphazardly to get around generic SSL/TLS problems when making normal API calls or scraping sites.

Since we only made a HEAD request, we’re just getting back headers, so let’s take a look at them:

str(httr::headers(res), 1)
## List of 18
##  $ date                     : chr "Fri, 10 Jul 2020 15:02:32 GMT"
##  $ content-type             : chr "text/html; charset=UTF-8"
##  $ content-length           : chr "11749"
##  $ connection               : chr "keep-alive"
##  $ etag                     : chr "\"7e0d5e2b6add\""
##  $ pragma                   : chr "no-cache"
##  $ cache-control            : chr "no-store, no-cache, must-revalidate, post-check=0, pre-check=0"
##  $ expires                  : chr "Thu, 19 Nov 1981 08:52:00 GMT"
##  $ x-frame-options          : chr "DENY"
##  $ set-cookie               : chr "PHPSESSID=bde5668131c14b765e3e75f8ed5514a0; path=/; secure; HttpOnly"
##  $ set-cookie               : chr "PHPSESSID=bde5668131c14b765e3e75f8ed5514a0; path=/; secure; HttpOnly"
##  $ set-cookie               : chr "PHPSESSID=bde5668131c14b765e3e75f8ed5514a0; path=/; secure; HttpOnly"
##  $ set-cookie               : chr "PHPSESSID=bde5668131c14b765e3e75f8ed5514a0; path=/; secure; HttpOnly"
##  $ set-cookie               : chr "PHPSESSID=bde5668131c14b765e3e75f8ed5514a0; path=/; samesite=lax; secure; httponly"
##  $ strict-transport-security: chr "max-age=31536000;"
##  $ x-xss-protection         : chr "1; mode=block;"
##  $ x-content-type-options   : chr "nosniff"
##  $ content-security-policy  : chr "default-src 'self'; script-src 'self' 'unsafe-inline'; img-src * data:; style-src 'self' 'unsafe-inline';"
##  - attr(*, "class")= chr [1:2] "insensitive" "list"

As an aside, I’ve always found the use of PHP code in security products quite, er, fascinating.

The value we’re really looking for here is etag (which really looks like ETag in the raw response).

Bishop Fox (and others) figured out that that header value contains a timestamp in the last 8 characters. That timestamp maps to the release date of the particular PAN-OS version. Since Palo Alto maintains multiple, supported versions of PAN-OS and generally releases patches for them all at the same time, the mapping to an exact version is not super precise, but it’s sufficient to get an idea of whether that system is at a current, supported patch level.

The last 8 characters of 7e0d5e2b6add are 5e2b6add, which — as Bishop Fox notes in their repo — is just a hexadecimal encoding of the POSIX timestamp, in this case, 1579903709 or 2020-01-24 22:08:29 GMT (we only care about the date, so really 2020-01-24).

We can compute that with R, but first we need to note that the value is surrounded by " quotes, so we’ll have to deal with that during the processing:

httr::headers(res) %>% 
  pluck("etag") %>% 
  gsub('"', '', .) %>% 
  substr(5, 12) %>% 
  as.hexmode() %>% 
  as.integer() %>% 
  anytime::anytime(tz = "GMT") %>% 
  as.Date() -> version_date

version_date
## [1] "2020-01-24"

To get the associated version(s), we need to look the date up in their table, which is in a fixed-width format that we can read via:

read_fwf(
  file = "https://raw.githubusercontent.com/noperator/panos-scanner/master/version-table.txt",
  col_positions = fwf_widths(c(10, 14), c("version", "date")),
  col_types = "cc",
  trim_ws = TRUE
) %>% 
  mutate(
    date = lubridate::mdy(date)
  ) -> panos_trans

panos_trans
## # A tibble: 153 x 2
##    version  date      
##    <chr>    <date>    
##  1 6.0.0    2013-12-23
##  2 6.0.1    2014-02-26
##  3 6.0.2    2014-04-18
##  4 6.0.3    2014-05-29
##  5 6.0.4    2014-07-30
##  6 6.0.5    2014-09-04
##  7 6.0.5-h3 2014-10-07
##  8 6.0.6    2014-10-07
##  9 6.0.7    2014-11-18
## 10 6.0.8    2015-01-13
## # … with 143 more rows

Now, let’s see what version or versions this might be:

filter(panos_trans, date == version_date)
## # A tibble: 2 x 2
##   version date      
##   <chr>   <date>    
## 1 9.0.6   2020-01-24
## 2 9.1.1   2020-01-24

Putting It All Together

We can make a command line script for this (example) scanner:

#!env Rscript
library(purrr)

gg <- glue::glue

# we also use {httr}, {readr}, {lubridate}, {anytime}, and {jsonlite}

args <- commandArgs(trailingOnly = TRUE)

stopifnot(
  c(
    "Must supply both IP address and port" = length(args) == 2
  )
)

ip <- args[1]
port <-  args[2]

httr::HEAD(
  url = gg("https://{ip}:{port}/global-protect/login.esp"),
  config = httr::config(
    ssl_verifyhost =FALSE, 
    ssl_verifypeer = FALSE
  )
) -> res

httr::headers(res) %>% 
  pluck("etag") %>% 
  gsub('"', '', .) %>% 
  substr(5, 12) %>% 
  as.hexmode() %>% 
  as.integer() %>% 
  anytime::anytime(tz = "GMT") %>% 
  as.Date() -> version_date

panos_trans <- readr::read_csv("panos-versions.txt", col_types = "cD")

res <- panos_trans[panos_trans[["date"]] == version_date,]

if (nrow(res) == 0) {
  cat(gg('{{"ip":"{ip}","port":"{port}","version"=null,"date"=null}}\n'))
} else {
  res$ip <- ip
  res$port <- port
  jsonlite::stream_out(res[,c("ip", "port", "version", "date")], verbose = FALSE)
}

Save that as panos-scanner.R and make it executable (provided you’re on a non-legacy operating system that doesn’t support such modern capabilities). Save panos_trans as a CSV file in the same directory and try it against another (sanitized IP/port) system:

./panos-scanner.R 10.20.30.40 5678                                                                                                                                                    1
{"ip":"10.20.30.40","port":"5678","version":"9.1.2","date":"2020-03-30"}

FIN

To be complete, the script should test all the URLs the ones in the script from Bishop Fox does and stand up many more guard rails to handle errors associated with unreachable hosts, getting headers from a system that is not a PAN-OS GP host, and ensuring the ETag is a valid date.

You can grab the code [from this repo](https://git.rud.is/hrbrmstr/2020-07-10-panos-rstats.

BIMI Up, Scotty! A look at Brand Indicators for Message Identification (BIMI) Adoption with R and the Alexa Top 1m

It seems that the need for MX, DKIM, SPF, and DMARC records for modern email setups were just not enough acronyms (and setup tasks) for some folks, resulting in the creation of yet-another-acronym — BIMI, or, Brand Indicators for Message Identification. The goal of BIMI is to “provide a mechanism for mail senders to publish a validated logotype that mail receivers can display with the senders’ messages.” You can read about the rationale for BIMI and the preliminary RFC for crafting BIMI DNS TXT records over a few caffeinated beverages. I’ll try to TL;DR the high points below.

The idea behind BIMI is to provide a visual indicator of the brand associated with a mail message; i.e. you’ll have an image to look at somewhere in the mail list display and/or mail message display of your mail client if it supports BIMI. This visual indicator is merely an image URL association with a brand mail domain through the use of a new special-prefix DNS TXT record. Mail intermediaries and mail clients are only supposed to allow presentation of BIMI-record provided images after verifying that the email domain itself conforms to the DMARC standard (which you should be using if you’re an organization/brand and shame on you if you’re not by now). In fact, the goal of BIMI is to help ensure:

the organization is legitimate
the domain names are controlled by the organization
the organization has current rights to display the indicator

When BIMI validation is being performed, the party requesting validation is currently authorized to do so by the organization and is who they say they are.

If you’re having flashbacks to the lost era of when SSL certificates were supposed to have similar integrity assertions, you’re not alone (thanks, LE).

What’s Really Going On?

I’m not part of any working group associated with BIMI, I just measure and study the internet for a living. As someone who is as likely to use alpine to peruse mail as I am a thick email client or (heaven forbid) web client, BIMI will be of little value to me since I’m not really going to see said images anyway.

Reading through all the BIMI (and associated) RFCs, email security & email marketing vendor blogs/papers, and general RFC commentators, BIMI isn’t solving any problem that well-armored DMARC configurations aren’t already solving. It appears to be driven mainly by brand marketing wonks who just want to shove brand logos in front of you and have one more way to track you.

Yep, tracking email perusals (even if it’s just a list view) will be one of the benefits (to brands and marketing firms) and is most assuredly a non-stated primary goal of this standard. To help illustrate this, let’s look at the BIMI record for one of the most notorious tracking brands on the planet, Verizon (in this case, Verizon Wireless). When you receive a BIMI-“enhanced” email from verizonwireless.com the infrastructure handling the email receipt will look for and process the BIMI header that was sent along for the ride and eventually query a TXT record for default._bimi.verizonwireless.com (or whatever the sender has specified instead of default — more on that in a bit). In this case the response will be:

v=BIMI1; l=https://ecrm.e.verizonwireless.com/AC/Global/Bling/Images/checkmark/verizon.svg;

which means the image they want displayed is at that URL. Your client will have to fetch that during an interactive session, so your IP address — at a minimum — will be leaked when that fetch happens.

Brands can specify something other than the default. selector with the email, so they could easily customize that to be a unique identifier which will “be you” and know when you’ve at least looked at said email in a list view (provided that’s how your email client will show it) if not in the email proper. Since this is a “high integrity” visual component of the message, it’s likely not going to be subject to the “do not load external images/content” rules you have setup (you do view emails with images turned off initially, right?).

So, this is likely just one more way the IETF RFC system is being abused by large corporations to continue to erode our privacy (and get their horribly designed logos in our faces).

Let’s see who are the early adopters of BIMI.

BIMI Through the Alexa Looking Glass

Amazon had stopped updating the Alexa Top 1m sites for a while but it’s been back for quite some time so we can use it to see how many sites in the top 1m have BIMI records.

We’ll use the {zdnsr} package (also on GitLab, SourceHut, BitBucket, and GitUgh) to perform a million default._bimi prefix queries and see how many valid BIMI TXT record responses we get.

library(zdnsr) # hrbrmstr/zdnsr on social coding sites
library(stringi)
library(urltools)
library(tidyverse)

refresh_publc_nameservers_list() # get a current list of active nameservers we can use

# read in the top1m
top1m <- read_csv("~/data/top-1m.csv", col_names = c("rank", "domain")) # http://s3.amazonaws.com/alexa-static/top-1m.csv.zip

# fire off a million queries, storing good results where we can pick them up later
zdns_query(
  entities = sprintf("%s.%s", "default._bimi", top1m$domain),
  query_type = "TXT",
  num_nameservers = 500,
  output_file = "~/data/top1m-bimi.json",
)

# ~10-30m later depending on your system/network/randomly chosen resolvers

bmi <- jsonlite::stream_in(file("~/data/top1m-bimi.json")) # using jsonlite vs ndjson since i don't want a "flat" structure

idx <- which(lengths(bmi$data$answers) > 0) # find all the ones with non-0 results

# start making a tidy data structure
tibble(
  answer = bmi$data$answers[idx]
) %>%
  unnest(answer) %>%
  filter(grepl("^v=BIM", answer)) %>% # only want BIMI records, more on this in a bit
  mutate(
    l = stri_match_first_regex(answer, "l=([^;]+)")[,2], # get the image link
    l_dom = domain(l) # get the image domain
  ) %>% 
  bind_cols(
    suffix_extract(.$name) # so we can get the apex domain below
  ) %>% 
  mutate(
    name_apex = glue::glue("{domain}.{suffix}"),
    name_stripped = stri_replace_first_regex(
      name, "^default\\._bimi\\.", ""
    )
  ) %>% 
  select(name, name_stripped, name_apex, l, l_dom, answer) -> bimi_df

Here’s what we get:

bimi_df
## # A tibble: 321 x 6
##    name       name_stripped  name_apex  l                            l_dom               answer                       
##    <chr>      <chr>          <glue>     <chr>                        <chr>               <chr>                        
##  1 default._… ebay.com       ebay.com   https://ir.ebaystatic.com/p… ir.ebaystatic.com   v=BIMI1; l=https://ir.ebayst…
##  2 default._… linkedin.com   linkedin.… https://media.licdn.com/med… media.licdn.com     v=BIMI1; l=https://media.lic…
##  3 default._… wish.com       wish.com   https://wish.com/static/img… wish.com            v=BIMI1; l=https://wish.com/…
##  4 default._… dropbox.com    dropbox.c… https://cfl.dropboxstatic.c… cfl.dropboxstatic.… v=BIMI1; l=https://cfl.dropb…
##  5 default._… spotify.com    spotify.c… https://message-editor.scdn… message-editor.scd… v=BIMI1; l=https://message-e…
##  6 default._… ebay.co.uk     ebay.co.uk https://ir.ebaystatic.com/p… ir.ebaystatic.com   v=BIMI1; l=https://ir.ebayst…
##  7 default._… asos.com       asos.com   https://content.asos-media.… content.asos-media… v=BIMI1; l=https://content.a…
##  8 default._… wix.com        wix.com    https://valimail-app-prod-u… valimail-app-prod-… v=BIMI1; l=https://valimail-…
##  9 default._… cnn.com        cnn.com    https://amplify.valimail.co… amplify.valimail.c… v=BIMI1; l=https://amplify.v…
## 10 default._… salesforce.com salesforc… https://c1.sfdcstatic.com/c… c1.sfdcstatic.com   v=BIMI1; l=https://c1.sfdcst…
## # … with 311 more rows

I should re-run this mass query since it usually takes 3-4 runs to get a fully comprehensive set of results (I should also really use work’s infrastructure to do the lookups against the authoritative nameservers for each organization like we do for our FDNS studies, but this was a spur-of-the-moment project idea to see if we should add BIMI to our studies and my servers are “free” whereas AWS nodes most certainly are not).

To account for the aforementioned “comprehensiveness” issues, we’ll round up the total from 310 to 400 (the average difference between 1 and 4 bulk queries is more like 5% than 20% but I’m in a generous mood), so 0.04% of the domains in the Alexa Top 1m have BIMI records. Not all of those domains are going to have MX records but it’s safe to say less than 1% of the brands on the Alexa Top 1m have been early BIMI adopters. This is not surprising since it’s not really a fully baked standard and no real clients support it yet (AOL doesn’t count, apologies to the Oathers). Google claims to be “on board” with BIMI, so once they adopt it, we should see that percentage go up.

Tracking isn’t limited to a tricked out dynamic DNS configuration that customizes selectors for each recipient. Since many brands use third party services for all things email, those clearinghouses are set to get some great data on you if these preliminary results are any indicator:

count(bimi_df, l_dom, sort=TRUE)
## # A tibble: 255 x 2
##    l_dom                                                                          n
##    <chr>                                                                      <int>
##  1 irepo.primecp.com                                                             13
##  2 www.letakomat.sk                                                               9
##  3 valimail-app-prod-us-west-2-auth-manager-assets.s3.us-west-2.amazonaws.com     8
##  4 static.mailkit.eu                                                              7
##  5 astatic.ccmbg.com                                                              5
##  6 def0a2r1nm3zw.cloudfront.net                                                   4
##  7 static.be2.com                                                                 4
##  8 www.christin-medium.com                                                        4
##  9 amplify.valimail.com                                                           3
## 10 bimi-host.250ok.com                                                            3
## # … with 245 more rows

The above code counted how many BIMI URLs are hosted at a particular domain and the top 5 are all involved in turning you into the product for other brands.

Speaking of brands, these are the logos of the early adopters which I made by generating some HTML from an R script and screen capturing the browser result:

FIN

The data from the successful BIMI results of the mass DNS query is at https://rud.is/dl/2020-02-21-bimi-responses.json.gz. Knowing there are results to be had, I’ll be setting up a regular (proper) mass-query of the Top 1m and see how things evolve over time and possibly get it on the work docket. We may just do a mass BIMI prefix query against all FDNS apex domains just to see a broader scale result, so stay tuned.

Drop note if you discover any more insights from the data (there are a few in there I’m saving for a future post) or your own BIMI inquiries; also drop a note if you have a good defense for BIMI other than marketing and tracking.

Davos 2020 World Economic Forum 2020 Global Risk Report Cyber Cliffs Notes

2020-01-21 – 09:14
Posted in Cybersecurity, Risk, Risk Assessment, Risk Management
Leave a Comment

Each year the World Economic Forum releases their Global Risk Report around the time of the annual Davos conference. This year’s report is out and below are notes on the “cyber” content to help others speed-read through those sections (in the event you don’t read the whole thing). Their expert panel is far from infallible, but IMO it’s worth taking the time to read through their summarized viewpoints. Some of your senior leadership are represented at Davos and either contributed to the report or will be briefed on the report, so it’s also a good idea just to keep an eye on what they’ll be told.

Direct link to report PDF: http://www3.weforum.org/docs/WEF_Global_Risk_Report_2020.pdf.

“Cyber” Cliffs Notes

Cyberattacks moved out of the Top 5 Global Risks in terms of Likelihood (page 2)
Cyberattacks remain in the upper-right risk quadrant (page 3)
Cyberattacks likelihood estimation reduced slightly but impact moved up a full half point to ~4.0 (out of 5.0) (page 4)
Cyberattacks are placed as directly related to named risks of: (page 5)
- information infrastructure breakdown, (76.2% of the 200+ member expert panel on short-term outlook)
- data fraud/theft, (75.0% of the 200+ member expert panel on short-term outlook) and
- adverse tech advances (<70% of the 200+ member expert panel on short-term outlook)
All three of which have their own relationships (it’s worth tracing them out as an exercise in downstream impact potential if one hasn’t worked through a risk relationship exercise before)
Cyberattacks remain on the long-term outlook (next 10 years) for both likelihood and impact by all panel sectors
Pages 61-71 cover the “Fourth Industrial Revolution” (4IR) and cyberattacks are mentioned on every page.
- There are 2025 market projections that might be useful as deck fodder.
- Interesting statistic that 50% of the world’s population is online and that one million additional people are joining the internet daily.
- The notion of nation-state mandated “parallel cyberspaces” is posited (we’re seeing that develop in Russia and some other countries right now).
- They also mention the proliferation of patents to create and enforce a first-mover advantage
- Last few pages of the section have a wealth of external resources that are worth perusing
In the health section on page 78 they mention the susceptibility of health data to cyberattacks
They list out specific scenarios in the back; many have a cyber component
- Page 92: “Geopolitical risk”: Interstate conflict with regional consequences — A bilateral or multilateral dispute between states that escalates into economic (e.g. trade/currency wars, resource nationalization), military, cyber, societal or other conflict.
- Page 92: “Technological risk”: Breakdown of critical information infrastructure and networks — Cyber dependency that increases vulnerability to outage of critical information infrastructure (e.g. internet, satellites) and networks, causing widespread disruption.
- Page 92: “Technological risk”: Large-scale cyberattacks — Large-scale cyberattacks or malware causing large economic damage, geopolitical tensions or widespread loss of trust in the internet.
- Page 92: “Technological risk”: Massive incident of data fraud or theft — Wrongful exploitation of private or official data that takes place on an unprecedented scale.

FIN

Hopefully this saved folks some time, and I’m curious as to how others view the Ouija board scrawls of this expert panel when it comes to cybersecurity predictions, scenarios, and ratings.