httpscan 🌐🔍

A configurable HTTP scanner for finding publicly available dumps, backups, configs, and more. It can send requests using other HTTP methods besides GET. It can bypass Cloudflare checks, but Node.js must be installed on your system.

I've been using the httpx tool for a long time. That tool is great except for its configurability (and it doesn't know how to bypass CloudFlare). I needed a declarative way to describe scanning rules, plus the tool had to download the found backups, dumps, etc., on its own. The same developers have nuclei, but it solves slightly different problems, and mainly, I can't just rewrite tons of code in Go, where doing many things is a pain.

Installation 🛠️

pipx install httpscan

Use pipx instead of pip to install packages containing executables.

Install the latest version from GitHub:

pipx install git+https://github.com/s3rgeym/httpscan.git

Usage 🚀

$ httpscan -i urls.txt -o results.json -vv --proxy 'socks5://localhost:1080'

Help:

$ httpscan -h

--workers specifies the total number of workers.
Each worker processes one link from the queue.
There is a --delay between attempts on the same link (site) because Nginx often limits the number of requests from one IP per second.
--parallel is the maximum number of parallel attempts for ALL PROBES run by different workers.

If the config path is not specified, files named httpscan.yml or httpscan.yaml will be searched in the current working directory or ~/.config.

Fields contained in the config: see Config.

The config has a probes field, which contains a list of ProbeDict objects.

Example probe for finding DB dumps:

probes:
# ...
- condition: status_code == 200
  match: INSERT INTO
  name: database dump
  path: /{db,dump,database,backup}.sql
  save_file: true

The repository contains a sample.httpscan.yml (you can move it to ~/.config/httpscan.yml).

Each element in the probes array contains a required name field with the name of the probe.
path is the path to be appended to each URL. The path supports brace expansion like in BASH, for example, /{foo,ba{r,z}}.biz (the paths /foo.biz, /bar.biz, and /baz.biz will be checked).
method specifies the HTTP method; params is for passing QUERY STRING parameters, data for parameters sent via application/x-www-form-urlencoded, json..., cookies..., headers...
condition allows filtering results. A built-in expression engine supports operators ==, !=, <, <=, >, >=, ! or NOT, AND or &&, OR or ||. =~ checks if a string matches a pattern, similar to calling bool(re.search(right, left)). All operators are case-insensitive. They can be grouped using parentheses. Available variables: status_code, content_length, content_type, title... For example, status_code == 200 && content_type == 'application/json'. Note that strings must be in quotes...
match, not_match check for matching a regular expression pattern in the server's response. extract and extract_all allow extracting content that matches a pattern. Since the response body can be huge, the probe reads the first 64 KB of data from the socket by default. For an HTML page, this is enough (according to this site, the average HTML size in 2022 was 31 KB), and archives can be checked for the absence of HTML tags.
save_file: true saves the file in case of success by default in the ./output/%hostname% directory.

Scanning results are output in JSONL format (JSON Lines, where each object is on a new line). Use jq to work with them.

{
  "content_languages": ["en"],
  "content_length": 303,
  "content_type": "application/octet-stream",
  "host": "domain.tld",
  "http_version": "1.1",
  "input": "https://domain.tld",
  "port": 443,
  "probe": {
    "name": "docker config file",
    "not_match": "^\\s*<[a-zA-Z]+",
    "path": "/{{prod,dev,}.env,Dockerfile{,.prod,.dev},docker-compose{,.prod,.dev}.yml}",
    "save_file": true
  },
  "response_headers": {
    "Accept-Ranges": "bytes",
    "Cache-Control": "no-cache, no-store, must-revalidate",
    "Content-Length": 303,
    "Date": "Tue, 16 Jul 2024 17:38:08 GMT",
    "Etag": "\"12f-60425670233e7\"",
    "Expires": "0",
    "Last-Modified": "Wed, 30 Aug 2023 15:15:48 GMT",
    "Pragma": "no-cache",
    "Server": "Apache",
    "Strict-Transport-Security": "max-age=31536000; includeSubDomains",
    "Vary": "User-Agent",
    "X-Content-Type-Options": "nosniff",
    "X-Frame-Options": "sameorigin",
    "X-XSS-Protection": "1; mode=block"
  },
  "response_url": "https://domain.tld/.env",
  "saved_as": "/tmp/x/domain.tld/.env",
  "saved_bytes": 303,
  "server": "Apache",
  "status_code": 200,
  "status_reason": "OK"
}

Other Features 🧩

Each link to be scanned uses a random User-Agent.
Proxy support, for example, socks5://localhost:1080.
--exclude-hosts can be used to pass a list of ignored hosts, and patterns with asterisks like *.shopify.com can be used to filter out subdomains. Domains can be written in any case.
Certain response codes can be skipped, for example, --exclude-statuses 401 403.

Development 🖥️

git clone ... && cd ...
python -m venv .venv
. .venv/bin/activate
# install all dependencies from pyproject.toml
pip install .

TODO

set_state: next_state
on_state: state_name
run_python: script_name

def run(...) -> ...:
    ...