diff --git a/docs/VULNERABILITY_CATALOG.md b/docs/VULNERABILITY_CATALOG.md index 5763d34..9615ca2 100644 --- a/docs/VULNERABILITY_CATALOG.md +++ b/docs/VULNERABILITY_CATALOG.md @@ -7,12 +7,12 @@ from each file's header comment, so this page cannot drift from the source. ## Totals -- **Test cases:** 62 -- **Expected detections:** 62 -- **`VULNERABLE:` markers:** 123 (individual lines a scanner should flag) -- **`SAFE:` markers:** 73 (lines a scanner must not flag — the false-positive control group) +- **Test cases:** 63 +- **Expected detections:** 63 +- **`VULNERABLE:` markers:** 124 (individual lines a scanner should flag) +- **`SAFE:` markers:** 74 (lines a scanner must not flag — the false-positive control group) - **Languages:** 8 — dotenv, go, java, javascript, json, python, ruby, text -- **CWE categories:** 46 — CWE-20, CWE-22, CWE-78, CWE-79, CWE-89, CWE-90, CWE-95, CWE-113, CWE-117, CWE-190, CWE-201, CWE-209, CWE-256, CWE-295, CWE-321, CWE-327, CWE-330, CWE-338, CWE-346, CWE-347, CWE-352, CWE-362, CWE-377, CWE-384, CWE-489, CWE-502, CWE-506, CWE-532, CWE-601, CWE-611, CWE-614, CWE-639, CWE-643, CWE-681, CWE-759, CWE-798, CWE-862, CWE-915, CWE-918, CWE-942, CWE-943, CWE-1236, CWE-1321, CWE-1333, CWE-1336, CWE-1357 +- **CWE categories:** 47 — CWE-20, CWE-22, CWE-78, CWE-79, CWE-89, CWE-90, CWE-95, CWE-113, CWE-117, CWE-185, CWE-190, CWE-201, CWE-209, CWE-256, CWE-295, CWE-321, CWE-327, CWE-330, CWE-338, CWE-346, CWE-347, CWE-352, CWE-362, CWE-377, CWE-384, CWE-489, CWE-502, CWE-506, CWE-532, CWE-601, CWE-611, CWE-614, CWE-639, CWE-643, CWE-681, CWE-759, CWE-798, CWE-862, CWE-915, CWE-918, CWE-942, CWE-943, CWE-1236, CWE-1321, CWE-1333, CWE-1336, CWE-1357 ## How coverage is scored @@ -36,6 +36,7 @@ counts as a detection. See `docs/SCANNER_INTEGRATION.md`. | Test case | File | CWE | Severity | Expected | Markers | |---|---|---|---|---|---| | Insecure deserialisation via ObjectInputStream | [`deserialization-object-input-stream.java`](../vulns/java/deserialization-object-input-stream.java) | CWE-502 | critical | yes | 2 vuln / 2 safe | +| Insecure regex pattern matching with user-controlled input | [`cwe-185-java.java`](../vulns/java/cwe-185-java.java) | CWE-185 | high | yes | 1 vuln / 1 safe | | SQL injection via Statement string concatenation | [`sqli-statement-concat.java`](../vulns/java/sqli-statement-concat.java) | CWE-89 | critical | yes | 2 vuln / 1 safe | | Reflected XSS via HttpServletResponse writer | [`xss-response-writer.java`](../vulns/java/xss-response-writer.java) | CWE-79 | high | yes | 2 vuln / 2 safe | | XXE via unconfigured DocumentBuilderFactory | [`xxe-document-builder.java`](../vulns/java/xxe-document-builder.java) | CWE-611 | high | yes | 2 vuln / 1 safe | diff --git a/vulns/VULNERABILITY_CATALOG.json b/vulns/VULNERABILITY_CATALOG.json index 153fdaf..0e41858 100644 --- a/vulns/VULNERABILITY_CATALOG.json +++ b/vulns/VULNERABILITY_CATALOG.json @@ -2,10 +2,10 @@ "schema": "threatcrush-testbed-catalog/1", "note": "Generated by scripts/generate-catalog.py \u2014 do not edit by hand.", "totals": { - "test_cases": 62, - "expected_detections": 62, - "vulnerable_markers": 123, - "safe_markers": 73, + "test_cases": 63, + "expected_detections": 63, + "vulnerable_markers": 124, + "safe_markers": 74, "languages": [ "dotenv", "go", @@ -35,6 +35,7 @@ "CWE-95", "CWE-113", "CWE-117", + "CWE-185", "CWE-190", "CWE-201", "CWE-209", @@ -175,6 +176,29 @@ 56 ] }, + { + "id": "java-insecure-regex-pattern-matching", + "file": "vulns/java/cwe-185-java.java", + "title": "Insecure regex pattern matching with user-controlled input", + "category": "java", + "language": "java", + "cwe": "CWE-185", + "cwes": [ + "CWE-185" + ], + "severity": "high", + "expected_detection": true, + "description": "User-controlled input is used to construct a regular expression", + "detection_target": "Pattern.compile or Pattern.matcher reached from untrusted", + "safe_guard": "Every payload is behind the always-false NEVER_RUN constant and", + "attribution": "line", + "vulnerable_lines": [ + 37 + ], + "safe_lines": [ + 47 + ] + }, { "id": "java-deserialization-object-input-stream", "file": "vulns/java/deserialization-object-input-stream.java", diff --git a/vulns/java/cwe-185-java.java b/vulns/java/cwe-185-java.java new file mode 100644 index 0000000..e672c92 --- /dev/null +++ b/vulns/java/cwe-185-java.java @@ -0,0 +1,54 @@ +/** + * @id java-insecure-regex-pattern-matching + * @test-case Insecure regex pattern matching with user-controlled input + * @cwe CWE-185 + * @severity high + * @language java + * @expected-detection true + * @description User-controlled input is used to construct a regular expression + * pattern that is then used for pattern matching. CWE-185 refers to + * incorrect regular expression syntax or usage, specifically when + * user input is used to build a regex pattern without proper escaping + * or validation. This can lead to ReDoS (Regular Expression Denial of + * Service) attacks or incorrect pattern matching logic, allowing + * attackers to bypass security controls or cause resource exhaustion. + * @safe-guard Every payload is behind the always-false NEVER_RUN constant and + * the repository has no Java build file, so nothing is compiled. No + * malicious regex pattern or input is included in this file. + * @detection-target Pattern.compile or Pattern.matcher reached from untrusted + * input without proper escaping or validation. + * + * NEVER RUN IN PRODUCTION — intentional test case for scanner validation. + */ +package vulns.java; + +import java.util.regex.Pattern; +import java.util.regex.Matcher; +import javax.servlet.http.HttpServletRequest; + +public class InsecureRegexPatternMatching { + + private static final boolean NEVER_RUN = false; + + public boolean validateInputVulnerable(HttpServletRequest req) { + if (NEVER_RUN) { + String userPattern = req.getParameter("pattern"); // SOURCE + String userInput = req.getParameter("input"); // SOURCE + Pattern pattern = Pattern.compile(userPattern); // VULNERABLE: CWE-185 + Matcher matcher = pattern.matcher(userInput); + return matcher.matches(); + } + return false; + } + + public boolean validateInputSafe(HttpServletRequest req) { + if (NEVER_RUN) { + String userInput = req.getParameter("input"); + // SAFE: hardcoded pattern with proper escaping, no user input in regex + Pattern pattern = Pattern.compile("^[a-zA-Z0-9]+$"); + Matcher matcher = pattern.matcher(userInput); + return matcher.matches(); + } + return false; + } +} \ No newline at end of file