/excavator

Excavator is a lightweight pure Golang git leak scanning tool based on SAP's credential digger (without machine learning models) which attempts to improve on performance by introducing parallel computation.

Primary LanguageGoApache License 2.0Apache-2.0

Excavator

Go Report Card License

Excavator is a lightweight pure Golang leak scanning tool which attempts to improve on performance by parallelising commit iteration.

CLI usage

Download a binary here.

# For scanning git repository (local or remote)
# Rules can be downloaded at resources/rules.yaml
excavator git <source> [flags]

# Dor scanning local directory
excavator fs <path> [flags]

Flags

  • -h , --help : display help
  • -c , --concurrent <int> : number of concurrent executions (defaults to 1), any integer given below 0 is considered as a single routine execution
  • -p , --path <string> : temporary local path to store the git repository (only applies to remote repository) (default .)
  • -r , --rules <string> : location of the rule declaration (defaults to resources/rules.yaml embedded in the binary)
  • -f , --format <string> : format of output result (default html) (currently supports yaml, html)

Global Flags

  • -v , -vv, -vvv : set verbosity levels

Scanning a repository without backend

excavator scan {repository}

Include in code

import (
  "github.com/ichbinfrog/excavator/pkg/scan"
)

func main() {
  c := &scan.GitScanner{}

  // Directory in which to store the cloned repository
  directory := ...
  
  // URL / local path of git repository
  // for private repositories the url can be set as
  // https://user:pass@host/repo.git
  repo := ...
  
  // path to rule file
  rule := ...

  // Number of concurrent go routines 
  concurrent := ...

  // Whether or not to show progress bar
  progressBar := ...

  // Output interface
  // Can be either
  //  - &YamlReport{}
  //  - &HTMLReport{}
  report := ...
  c.New(repo, directory, rule, report, progressBar)
}

Declaring rules

# rules.yaml
#
apiVersion: v1
rules:
  - # regex of rule
    definition: EAACEdEose0cBA[0-9A-Za-z]+
    # category of rule
    category: token
    # description (optional)
    description: facebook access token rule

# list of regexes of file to ignore
black_list:
  - '.*sample.*'

# list of parsers
# parsers are rules that require additional context for analysing
# for potential leaks with more precision
#
# currently supports "env" and "dockerfile" 
parsers:
  - type: "env" 
    extensions:
      - ".env" 

    # the parser uses theses values to check if the key in the <key> = <value>
    # form contains potential leaks 
    keys:               
      - "pass"
      - "host"
      - "proxy"
      - "key"

  - type: "dockerfile"
    extensions:
      - "Dockerfile"
    # keys defaults to 
    # ["pass", "host", "proxy", "key"] if not defined

# Whether or not to explore files that are in archives
# e.g. tar, gzip, zip, rar...
compressed: True