Create query
To use Snyk Code custom rules to create queries with suggestive AI support, you can choose from the provided templates and predicates. Alternatively, you can create your own predicates and save them as a custom rule.
Consider the following query examples and rules to use with Snyk Code custom rules. A CWE 312 query example is provided on this page.
Simple syntactical query
Copy the following source code snippet in the snippet window and select C# as the language
It is only a snippet and not a full program. It will not compile.
Running the query
Enter the following queries in the query window and press Run Query to see the results.
Select
body
by using the query:“body”
This query does not select the Body with a capital B. The query language is case-sensitive.
Add
Body
to the findings so the query becomesOr<”body”,”Body”>
.You can achieve the same outcome using a regex
~"body|Body"
or~"[Bb]ody"
Do something more complex regex and query:
~"[a-z0-9!#$%&'*+/=?^_{|}~-]+(?:\.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*@(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?"
It matches the hardcoded email address.
Try it yourself
Run the following query over your code ~"([a-zA-Z0-9+/]{40})"
If you find something, check it out first, as you might leak your AWS secrets.
If you are interested in a certain type of object, you can use templates. For example, the query CallExpression<"Format">
matches a function call or Literal<"nobody@notrealdomain.co.uk">
matches the string with the email address.
A data flow or taint analysis
For this example, a JavaScript code snippet is used. You can copy it in the snippet window and select JavaScript.
Snyk Code knows a list of possible sources of external data in the predicate PRED:AnySource
. The following query shows you that app.post()
is identified.
Query PRED:SqliSinks
shows you that query()
is part of that list of SQL injection sinks. The query engine comes with many different predicates for various source, sink, and sanitizer types. Check the list of predicates to see them all.
To check whether the data flows into a SQL injection sink, use the following: DataFlowsInto<PRED:SqliSink>
. It shows you that in the program, data from the req
parameter flow into query()
taking several turns.
If the data flow is also going through a sanitizer, you can use a specialized template. Change the query to Taint<PRED:AnySource, PRED:SqliSanitizer, PRED:SqliSink>
There is nothing language-specific in the query. It would work on similar code in other languages.
Net new data flow rule
Create a new rule because Snyk is not aware of the proprietary source built in-house, resulting in missed findings.
Use a data flow template known as Taint
when creating a data flow query.
You can configure the following parameters:
Source: The first parameter indicates where the data flow starts.
Sanitizer: The second parameter indicates a known sanitizer that would sanitize the data, resulting in the data's not being tainted
Sink: The third parameter indicating where the data flow ends
Custom predicates are indicated by writing their names within brackets. In this scenario, the custom method is called SourceFoo
.
With this query, you can look for the data flow that originates in SourceFoo
. A source unknown to Snyk ends up in a known vulnerable cross-site scripting (XSS) Sink and does not pass through a known cross-site scripting (XSS) Sanitizer. Therefore, the assumption is that the data is tainted.
Extend a data flow rule
Recreate a Snyk rule and add a source to the current Snyk known vulnerable source list because they are not being taken into account in the scans, resulting in missed vulnerabilities.
Like the Net new data flow rule, the Taint
data flow template is used with an Or
operator. Operators are available to create logical statements for your queries, such as Or
or And
.
Run the data flow rule using both the Snyk known sources but also a custom source called .
With this query, you look for the data flow that originates in a Snyk known source OR “SourceFoo
” . A source unknown to Snyk ends up in a known vulnerable cross-site scripting (XSS) Sink and does not pass through a known cross-site scripting (XSS) Sanitizer. Therefore, the assumption is that the data is tainted.
Any statement that uses an operator will be written within angle brackets < statement >
.
Context added to data flow rule
Recreate a Snyk rule and remove a source from the current Snyk known vulnerable sources because this source is not vulnerable within the context of an application.
Like the Net new data flow and Extend a data flow rules, the Taint
data flow template is used with an And
operator. A declarative negative statement (Not
) is used to indicate the false case of the statement and not the true case.
Run the data flow rule using the Snyk known sources, removing SnykSource
from the results. In this example, SnykSource
is a Snyk-known source that is used within the regular general AnySource
predicate.
With this query, you look for the data flow that originates in a known Snyk source but remove results that come from SnykSource
that end up in a known vulnerable cross-site scripting (XSS) Sink and do not pass through a known cross-site scripting (XSS) Sanitizer. Therefore, the assumption is that the data is tainted.
High recall mode
See all the sources and sinks in the source code to understand every location where data can flow from or to. This analysis is conducted regardless of the presence of data flows, allowing users to assess coverage comprehensively.
This mode is often used as an investigatory tool to gain deeper insights into the code stack and to comprehend where data originates and terminates within the application.
To initiate this analysis, use the AnySource
or AnySink
predicates, which encompass all Snyk-known sources and sinks, respectively.
This query identifies and highlights every known source within the code being analyzed.
Similarly, this query identifies and highlights every known sink within the code, providing a complete view of potential data endpoints.
Assessing coverage and identifying missing data flow links
Identifying gaps in data flow paths is crucial for understanding the destinations of user data, especially if it does not end up in expected locations such as databases or file systems. These gaps may reveal the use of unsupported libraries or frameworks—or components thereof—potentially leading to false negatives. This insight is essential for comprehensive security assessments and ensuring robust coverage.