Skip to content

    Type a page, a command such as keryn scan, or a rule id.

    general-injection-kotlin-v1

    View as Markdown

    12 rules · Keryn Software Licence · rules 0.8.0 · all packs

    How to read a rule
    Severity
    How bad the finding is if it is real: error fails a default gate, warning and info do not.
    Confidence
    P0 exploitable as written; P1 dangerous in context; P2 a heuristic or a posture check: confirm by reading the code.
    False-positive risk
    low: an exact API or flag, so hits are almost always real; medium: the pattern is legitimate in many programs, so check each hit; high: it finds a lead to confirm, so expect benign hits.
    Edge
    The relationship the finding sits on. reaches: content or traffic gets to a thing; calls: one thing invokes another; reads: one thing reads from another; writes: one thing writes to another; sends: data leaves for another party; grants: authority is given or left unchecked; depends_on: something relies on a component it does not verify; trusts: one party believes another without evidence. The boundary it crosses, when there is one: user->prompt (untrusted input reaching a prompt); model->code (model output reaching code, a shell, a query or a page); secret->outside (a secret leaving the trusted side); supply-chain (code, models or prompts pulled in from outside).
    Taxonomy chips
    References for readers who work in OWASP LLM 2025, OWASP Agentic, CWE or NIST AI RMF. They say where a rule belongs, not that a framework is satisfied.

    keryn.general.kotlin.path-traversal-request-to-file

    Section titled “keryn.general.kotlin.path-traversal-request-to-file”

    error kotlin data-flow rule kotlin, ktor, spring, android

    A request value reaches a file read, write, delete or file response without being reduced to a file name or checked against the base directory. A value such as ../../etc/passwd escapes the intended folder. Use File(x).name, or normalize and verify startsWith(base) / canonicalPath before opening.

    Confidence P1: dangerous in context. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: reaches content or traffic gets to a thing; it crosses the user->prompt boundary (untrusted input reaching a prompt).

    CWE-22 LLM05

    Fix: Reduce the value to File(x).name, or build the File, take canonicalFile (or normalize()), and refuse the request unless it startsWith the canonical base directory. Prefer an id resolved through a lookup instead of a path.

    References: https://ktor.io/docs/server-requests.html · https://kotlinlang.org/api/core/kotlin-stdlib/kotlin.io/read-text.html

    error kotlin data-flow rule kotlin, java-util-zip, android

    An archive entry name is used to build the output file of an extraction. A crafted entry such as ../../x writes outside the target directory (Zip Slip). Resolve the entry against the destination, take its canonical path and refuse anything that does not start with the destination.

    Confidence P1: dangerous in context. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: reads one thing reads from another.

    CWE-22 LLM05

    Fix: val target = File(dest, entry.name).canonicalFile; require(target.startsWith(dest.canonicalFile)) before outputStream() / FileOutputStream(target).

    References: https://docs.oracle.com/en/java/javase/21/docs/api/java.base/java/util/zip/ZipEntry.html#getName()

    keryn.general.kotlin.sql-injection-request-to-query

    Section titled “keryn.general.kotlin.sql-injection-request-to-query”

    error kotlin data-flow rule kotlin, ktor, spring, jdbc, android-sqlite, exposed

    A request, intent or deep-link value is interpolated or concatenated into SQL text executed through JDBC, JdbcTemplate, JPA or Android SQLite. The value can change the query. Keep the SQL constant with ? placeholders and bind values (PreparedStatement.setX, selectionArgs, query parameters).

    Confidence P1: dangerous in context. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: calls one thing invokes another; it crosses the model->code boundary (model output reaching code, a shell, a query or a page).

    CWE-89 LLM05

    Fix: Write the SQL as a constant string with ? or :name placeholders and bind values separately (prepareStatement + setString, rawQuery(sql, arrayOf(x)), JdbcTemplate args). Map column or table names to an allow-list.

    References: https://developer.android.com/reference/android/database/sqlite/SQLiteDatabase#rawQuery(java.lang.String,%20java.lang.String) · https://docs.oracle.com/en/java/javase/21/docs/api/java.sql/java/sql/PreparedStatement.html

    keryn.general.kotlin.command-injection-request-to-shell

    Section titled “keryn.general.kotlin.command-injection-request-to-shell”

    error kotlin data-flow rule kotlin, ktor, spring, android

    A request value is run as an operating system command, either as the program itself or as the command string of a shell (sh -c). The value can append further commands. Run a fixed program with each argument as a separate list element and validate arguments against an allow-list.

    Confidence P1: dangerous in context. False-positive risk low: an exact API or flag, so hits are almost always real.

    Edge inspected: calls one thing invokes another; it crosses the model->code boundary (model output reaching code, a shell, a query or a page).

    CWE-78 LLM05

    Fix: Use ProcessBuilder(“tool”, “–flag”, arg) or listOf(…) with a constant executable; do not pass input through sh -c or cmd /c. Validate arguments against a strict regex or allow-list.

    References: https://docs.oracle.com/en/java/javase/21/docs/api/java.base/java/lang/ProcessBuilder.html

    keryn.general.kotlin.ssrf-request-to-http-client

    Section titled “keryn.general.kotlin.ssrf-request-to-http-client”

    warning kotlin data-flow rule kotlin, ktor-client, spring, okhttp

    A request value decides which URL the server fetches. An attacker can reach internal services, cloud metadata or loopback. Parse the URL, require https, compare the host to an allow-list and reject private and loopback addresses before the request.

    Confidence P1: dangerous in context. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: reaches content or traffic gets to a thing; it crosses the user->prompt boundary (untrusted input reaching a prompt).

    CWE-918 LLM05

    Fix: Do not pass a user supplied string as a request URL. Accept an id mapped to a fixed URL, or parse it, require https, check the host against an allow-list, resolve and reject non-public IPs, and disable or re-validate redirects.

    References: https://ktor.io/docs/client-requests.html · https://docs.spring.io/spring-framework/reference/integration/rest-clients.html

    keryn.general.kotlin.xss-request-to-html-response

    Section titled “keryn.general.kotlin.xss-request-to-html-response”

    warning kotlin data-flow rule kotlin, ktor, spring, servlet

    A request value is written into an HTML response body without encoding (Ktor respondText with text/html, or a servlet writer). The value runs as markup or script in the visitor’s browser. Encode with HtmlUtils.htmlEscape, the OWASP Encoder, or build the page with kotlinx.html / a template engine that escapes.

    Confidence P1: dangerous in context. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: calls one thing invokes another; it crosses the model->code boundary (model output reaching code, a shell, a query or a page).

    CWE-79 LLM05

    Fix: Escape request-derived text for the HTML context (HtmlUtils.htmlEscape, Encode.forHtml) or render with kotlinx.html, which escapes text nodes by default. Return text/plain when no markup is intended.

    References: https://ktor.io/docs/server-html-dsl.html · https://docs.spring.io/spring-framework/docs/current/javadoc-api/org/springframework/web/util/HtmlUtils.html

    keryn.general.kotlin.csrf-protection-disabled

    Section titled “keryn.general.kotlin.csrf-protection-disabled”

    warning kotlin kotlin, spring-security

    Spring Security CSRF protection is switched off for the whole filter chain. Cookie or session authenticated state-changing requests can then be forged from another site. Keep CSRF enabled, or disable it only for a stateless token-authenticated API and say so.

    Confidence P2: a heuristic or a posture check: confirm by reading the code. False-positive risk high: it finds a lead to confirm, so expect benign hits.

    Edge inspected: calls one thing invokes another.

    CWE-352 LLM05

    Fix: Leave csrf enabled for browser sessions (CookieCsrfTokenRepository.withHttpOnlyFalse() for SPAs). For a pure bearer-token API, restrict the exemption with ignoringRequestMatchers for those paths instead of disabling it globally.

    References: https://docs.spring.io/spring-security/reference/servlet/exploits/csrf.html

    keryn.general.kotlin.sql-built-from-variable

    Section titled “keryn.general.kotlin.sql-built-from-variable”

    warning kotlin kotlin, jdbc, spring, jpa, hibernate, android, exposed

    The SQL text passed to this call is assembled from a variable by concatenation, a string template (“$x” / “${x}”) or format(), so whatever the variable holds becomes part of the statement. Keep the SQL constant and bind values as ? or named parameters instead.

    Confidence P2: a heuristic or a posture check: confirm by reading the code. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: calls one thing invokes another; it crosses the model->code boundary (model output reaching code, a shell, a query or a page).

    CWE-89 LLM05

    Fix: Write the SQL as a constant string with ? or :name placeholders and bind values through PreparedStatement.setX, JdbcTemplate arguments, or Query.setParameter. Table and column names cannot be bound: map them from an allow-list. Numeric values converted with toInt() / toLong() are not flagged.

    References: https://docs.oracle.com/en/java/javase/21/docs/api/java.sql/java/sql/PreparedStatement.html · https://kotlinlang.org/docs/strings.html#string-templates · https://cheatsheetseries.owasp.org/cheatsheets/SQL_Injection_Prevention_Cheat_Sheet.html

    keryn.general.kotlin.insecure-deserialization-object-input-stream

    Section titled “keryn.general.kotlin.insecure-deserialization-object-input-stream”

    warning kotlin kotlin, java-serialization

    A java.io.ObjectInputStream is read with readObject/readUnshared and no ObjectInputFilter or serial filter is visible. If the bytes come from a request, socket, queue or file another party can write, they can instantiate gadget classes. Install an ObjectInputFilter (setObjectInputFilter / setSerialFilter), or use a data format.

    Confidence P2: a heuristic or a posture check: confirm by reading the code. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: depends_on something relies on a component it does not verify; it crosses the supply-chain boundary (code, models or prompts pulled in from outside).

    CWE-502 LLM03

    Fix: Prefer kotlinx.serialization or JSON over Java serialization. When it must stay, call ois.setObjectInputFilter(ObjectInputFilter.Config.createFilter(“com.example.*;!*”)) before the first read.

    References: https://docs.oracle.com/en/java/javase/21/core/serialization-filtering1.html

    keryn.general.kotlin.insecure-deserialization-xstream

    Section titled “keryn.general.kotlin.insecure-deserialization-xstream”

    warning kotlin kotlin, xstream

    XStream.fromXML is called on an XStream instance with no allowTypes / allowTypesByWildcard / addPermission or setupDefaultSecurity visible. Untrusted XML can then instantiate arbitrary classes. Configure a type allow-list before parsing.

    Confidence P2: a heuristic or a posture check: confirm by reading the code. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: depends_on something relies on a component it does not verify; it crosses the supply-chain boundary (code, models or prompts pulled in from outside).

    CWE-502 LLM03

    Fix: Call XStream.setupDefaultSecurity(xstream) and xstream.allowTypes(…) / allowTypesByWildcard before the first fromXML. Never feed untrusted XML to an XStream with the default (permissive) security.

    References: https://x-stream.github.io/security.html

    keryn.general.kotlin.insecure-deserialization-snakeyaml

    Section titled “keryn.general.kotlin.insecure-deserialization-snakeyaml”

    warning kotlin kotlin, snakeyaml

    A SnakeYAML Yaml built with the default constructor loads a document. Before SnakeYAML 2.0 the default Constructor instantiates any class named by a global tag (!!), so untrusted YAML can run code. Construct it with Yaml(SafeConstructor(LoaderOptions())) or upgrade and restrict tags with a TagInspector.

    Confidence P2: a heuristic or a posture check: confirm by reading the code. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: depends_on something relies on a component it does not verify; it crosses the supply-chain boundary (code, models or prompts pulled in from outside).

    CWE-502 LLM03

    Fix: Use Yaml(SafeConstructor(LoaderOptions())) for untrusted input, or SnakeYAML 2.x with a LoaderOptions TagInspector that allows only the expected types.

    References: https://bitbucket.org/snakeyaml/snakeyaml/wiki/Documentation

    keryn.general.kotlin.insecure-deserialization-jackson-default-typing

    Section titled “keryn.general.kotlin.insecure-deserialization-jackson-default-typing”

    warning kotlin kotlin, jackson

    Jackson default typing is switched on. Class names in the JSON then choose the instantiated type, which is a deserialization gadget vector for untrusted input. Use @JsonTypeInfo on specific base types, or pass a restrictive PolymorphicTypeValidator built with BasicPolymorphicTypeValidator.builder().allowIfSubType(…).

    Confidence P2: a heuristic or a posture check: confirm by reading the code. False-positive risk medium: the pattern is legitimate in many programs, so check each hit.

    Edge inspected: depends_on something relies on a component it does not verify; it crosses the supply-chain boundary (code, models or prompts pulled in from outside).

    CWE-502 LLM03

    Fix: Remove enableDefaultTyping. If polymorphism is required, call activateDefaultTyping with a BasicPolymorphicTypeValidator that allows only your own base package, never LaissezFaireSubTypeValidator.

    References: https://github.com/FasterXML/jackson-docs/wiki/JacksonPolymorphicDeserialization