18 releases (6 breaking)

new 0.11.3 Jun 23, 2022
0.10.1 Jun 8, 2022
0.1.4 Jan 28, 2022

#338 in Command line utilities

Download history 1/week @ 2022-03-09 18/week @ 2022-03-16 8/week @ 2022-03-23 4/week @ 2022-04-06 20/week @ 2022-04-27 26/week @ 2022-05-04 451/week @ 2022-05-11 97/week @ 2022-05-18 36/week @ 2022-05-25 245/week @ 2022-06-01 92/week @ 2022-06-08 86/week @ 2022-06-15 57/week @ 2022-06-22

493 downloads per month

GPL-3.0 license

69KB
1.5K SLoC

dionysos

Scanner for various IoCs

Installation

sudo apt install libyara-dev
cargo install dionysos

Features

Feature Details
Scanners filenames (by regular expressions), similar filenames (Levenshtein), yara, hashes
Output formats human-readable text (txt), comma-separated values (csv, conforming to RFC4180), JavaScript Object Notation (json), can be selected with --format <txt|csv|json>
Scan of compressed files yara-scan of gz and bz2 compressed files is supported; see -C switch. Be aware that files are decompressed into a decompression buffer, and that every thread gets its own decompression buffer. You should make sure that you have sufficient memory. If you need larger buffers, you can limit the number of threads using --threads
Special features yara-scan in Windows evtx files and Windows registry hives using --evtx and --reg

Usage

dionysos 0.11.1
Jan Starke <Jan.Starke@t-systems.com>
Scanner for various IoCs

USAGE:
    dionysos [OPTIONS]

OPTIONS:
    -C, --scan-compressed
            allow yara to scan compressed files. Currently, xz, bz2 and gz are supported

        --decompression-buffer <DECOMPRESSION_BUFFER_SIZE>
            maximum size (in MiB) of decompression buffer (per thread), which is used to scan
            compressed files [default: 128]

        --evtx
            also do YARA scan in Windows EVTX records (exported as JSON)

    -f, --format <OUTPUT_FORMAT>
            output format [default: txt] [possible values: csv, txt]

    -F, --filename <FILENAMES>
            regular expression to match against the basename of files. This parameter can be
            specified multiple times

    -h, --help
            Print help information

    -H, --file-hash <FILE_HASH>
            Hash of file to match against. Use any of MD5, SHA1 or SHA256

    -L, --log-file <LOG_FILE>
            path of the file to write logs to. Logs will always be appended

        --omit-levenshtein
            do not run the Levenshtein scanner

    -p, --threads <THREADS>
            use the specified NUMBER of threads [default: 16]

    -P, --path <PATH>
            path which must be scanned

        --progress
            display a progress bar (requires counting the number of files to be scanned before a
            progress bar can be displayed)

    -q, --quiet
            Less output per occurrence

        --reg
            also do YARA scan in Windows registry hive files

    -s, --print-strings
            print matching strings (only used by yara currently)

    -v, --verbose
            More output per occurrence

    -V, --version
            Print version information

    -Y, --yara <YARA>
            use yara scanner with the specified ruleset. This can be a single file, a zip file or a
            directory containing lots of yara files. Yara files must end with 'yar' or 'yara', and
            zip files must end with 'zip'

        --yara-timeout <YARA_TIMEOUT>
            timeout for the yara scanner, in seconds [default: 240]

Developer guide

How to add scanners

1. Implement a special result type for the scanner

For example, say we want to scan for files whose name match a regular expression. Our finding type could look like this:

struct FilenameFinding {
    filename: String,
    pattern: regex::Regex,
}

Every finding type needs to implement Display and ScannerFinding:

impl ScannerFinding for FilenameFinding {

    fn format_readable(&self, file: &str, _show_details: bool) -> Vec<String> {
        vec![
            format!("the name of '{}' matches the pattern /{}/", file, self.pattern)
        ]
    }

    fn format_csv<'a, 'b>(&'b self, file: &'a str) -> HashSet<crate::scanner_result::CsvLine> {
        hashset![CsvLine::new("Filename", &self.filename, file, String::new())]
    }

    fn to_json(&self, file: &str) -> serde_json::Value {
        json!({
            "01_scanner": "filename",
            "02_suspicious_file": file,
            "03_pattern": format!("{}", self.pattern)
        })
    }
}

2. Implementation of the scanner

Take, for example, the FilenameScanner, which tries to do a simple filename match:

pub struct FilenameScanner {
    patterns: Vec<regex::Regex>,
}

impl FilenameScanner {
    pub fn new(patterns: Vec<regex::Regex>) -> Self {
        Self {   
            patterns,
        }
    }
}

impl Display for FilenameScanner {
    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
        write!(f, "{}", "FilenameScanner")
    }
}

impl FileScanner for FilenameScanner
{
    fn scan_file(&self, file: &DirEntry) -> Vec<anyhow::Result<Box<dyn ScannerFinding>>> {
        let file = file.path();
        let filename = file.to_str().unwrap();
        let mut results = Vec::new();
        for pattern in self.patterns.iter() {
            if pattern.is_match(&filename) {
                results.push(
                    Ok(
                        Box::new(
                            FilenameFinding{
                                filename: filename.to_owned(),
                                pattern: pattern.clone()
                            }
                        ) as Box<dyn ScannerFinding>
                    )
                )
            }
        }
        results
    }
}

3. Add your scanner to the scanner chain

Which is currently hard-coded in Dionysos::run() (in src/dionysos.rs)

Dependencies

~14MB
~310K SLoC