This scanner can flag anti-patterns found in a codebase or require that certain strings be present. This might be useful for preventing the use of dangerous methods like eval()
in Ruby (which might allow for RCE) or dangerouslySetInnerHTML
in React (which might allow for XSS). By default, all found patterns are added to the info section of the report. If a found pattern is forbidden, this scanner will fail and the message
will be show to the developer in the report to give additional context on why this was an issue. A required
pattern must be found in order for the scan to pass.
The tool sift, written in Go, is used to perform the pattern matching.
scanner_configs:
PatternSearch:
matches:
- regex: dangerouslySetInnerHTML
message: Do not use dangerouslySetInnerHTML to render user controlled input.
forbidden: true
exclude_directory:
- node_modules
exclude_extension:
- md
- regex: "# Thread Model"
message: All repos must contain a documented threat model.
required: true
exclude_extension:
- .rb
- .js