| 1 | package org.egothor.methodatlas.ai; | |
| 2 | ||
| 3 | /** | |
| 4 | * Provides the optimized built-in taxonomy used to guide AI-based security | |
| 5 | * classification when prompt compactness and model reliability are prioritized. | |
| 6 | * | |
| 7 | * <p> | |
| 8 | * This class supplies a condensed taxonomy definition intended for use with | |
| 9 | * {@link org.egothor.methodatlas.ai.AiOptions.TaxonomyMode#OPTIMIZED}. In | |
| 10 | * contrast to {@link DefaultSecurityTaxonomy}, this variant is structured to | |
| 11 | * improve AI classification consistency by reducing prompt verbosity while | |
| 12 | * preserving the same controlled category set and classification intent. | |
| 13 | * </p> | |
| 14 | * | |
| 15 | * <h2>Design Goals</h2> | |
| 16 | * | |
| 17 | * <ul> | |
| 18 | * <li>minimize prompt length without changing the supported taxonomy</li> | |
| 19 | * <li>increase deterministic model behavior</li> | |
| 20 | * <li>reduce ambiguity in category selection</li> | |
| 21 | * <li>preserve professional terminology and decision rules</li> | |
| 22 | * </ul> | |
| 23 | * | |
| 24 | * <p> | |
| 25 | * The taxonomy text returned by this class is intended to be embedded directly | |
| 26 | * into AI prompts and therefore favors concise, machine-oriented instruction | |
| 27 | * structure over explanatory prose. | |
| 28 | * </p> | |
| 29 | * | |
| 30 | * <p> | |
| 31 | * This class is a non-instantiable utility holder. | |
| 32 | * </p> | |
| 33 | * | |
| 34 | * @see DefaultSecurityTaxonomy | |
| 35 | * @see org.egothor.methodatlas.ai.AiSuggestionEngineImpl | |
| 36 | * @see org.egothor.methodatlas.ai.AiOptions.TaxonomyMode | |
| 37 | */ | |
| 38 | public final class OptimizedSecurityTaxonomy { | |
| 39 | /** | |
| 40 | * Prevents instantiation of this utility class. | |
| 41 | */ | |
| 42 | private OptimizedSecurityTaxonomy() { | |
| 43 | } | |
| 44 | ||
| 45 | /** | |
| 46 | * Returns the optimized built-in taxonomy text used for AI classification. | |
| 47 | * | |
| 48 | * <p> | |
| 49 | * The returned taxonomy is a compact instruction set designed for large | |
| 50 | * language models performing security classification of JUnit test methods. It | |
| 51 | * preserves the same controlled tag set as the default taxonomy while | |
| 52 | * presenting the rules in a shorter, more model-oriented structure. | |
| 53 | * </p> | |
| 54 | * | |
| 55 | * <p> | |
| 56 | * The taxonomy defines: | |
| 57 | * </p> | |
| 58 | * <ul> | |
| 59 | * <li>the meaning of a security-relevant test</li> | |
| 60 | * <li>the mandatory {@code security} umbrella tag</li> | |
| 61 | * <li>the allowed category tags</li> | |
| 62 | * <li>selection rules for assigning taxonomy tags</li> | |
| 63 | * <li>guidance for use of the optional {@code owasp} tag</li> | |
| 64 | * <li>the required {@code SECURITY: <property> - <scenario>} display name | |
| 65 | * format</li> | |
| 66 | * </ul> | |
| 67 | * | |
| 68 | * <p> | |
| 69 | * This optimized variant is suitable when improved model consistency or shorter | |
| 70 | * prompt size is more important than human-oriented explanatory wording. | |
| 71 | * </p> | |
| 72 | * | |
| 73 | * @return optimized taxonomy text used to instruct AI classification | |
| 74 | * | |
| 75 | * @see DefaultSecurityTaxonomy#text() | |
| 76 | * @see org.egothor.methodatlas.ai.AiSuggestionEngineImpl | |
| 77 | */ | |
| 78 | public static String text() { | |
| 79 |
1
1. text : replaced return value with "" for org/egothor/methodatlas/ai/OptimizedSecurityTaxonomy::text → SURVIVED |
return """ |
| 80 | SECURITY TEST CLASSIFICATION SPECIFICATION | |
| 81 | ========================================== | |
| 82 | ||
| 83 | Goal | |
| 84 | ---- | |
| 85 | ||
| 86 | Classify JUnit 5 test methods that validate security properties. | |
| 87 | ||
| 88 | The output MUST follow the allowed tag taxonomy and MUST NOT introduce new tags. | |
| 89 | ||
| 90 | ||
| 91 | Security-Relevant Test Definition | |
| 92 | --------------------------------- | |
| 93 | ||
| 94 | A test is security-relevant when it verifies any of the following: | |
| 95 | ||
| 96 | • authentication behavior | |
| 97 | • authorization decisions | |
| 98 | • cryptographic correctness | |
| 99 | • validation of untrusted input | |
| 100 | • protection against injection attacks | |
| 101 | • protection of sensitive data | |
| 102 | • security event logging | |
| 103 | • secure error handling | |
| 104 | ||
| 105 | If failure of the test could allow: | |
| 106 | ||
| 107 | • unauthorized access | |
| 108 | • data exposure | |
| 109 | • privilege escalation | |
| 110 | • security control bypass | |
| 111 | ||
| 112 | then the test is security-relevant. | |
| 113 | ||
| 114 | ||
| 115 | Mandatory Tag | |
| 116 | ------------- | |
| 117 | ||
| 118 | Every security-relevant test MUST contain: | |
| 119 | ||
| 120 | security | |
| 121 | ||
| 122 | ||
| 123 | Allowed Category Tags | |
| 124 | --------------------- | |
| 125 | ||
| 126 | Only the following tags are permitted: | |
| 127 | ||
| 128 | auth | |
| 129 | access-control | |
| 130 | crypto | |
| 131 | input-validation | |
| 132 | injection | |
| 133 | data-protection | |
| 134 | logging | |
| 135 | error-handling | |
| 136 | owasp | |
| 137 | ||
| 138 | ||
| 139 | Category Semantics | |
| 140 | ------------------ | |
| 141 | ||
| 142 | auth | |
| 143 | authentication validation | |
| 144 | identity verification | |
| 145 | credential checks | |
| 146 | token/session validation | |
| 147 | ||
| 148 | access-control | |
| 149 | authorization enforcement | |
| 150 | permission checks | |
| 151 | role evaluation | |
| 152 | ownership validation | |
| 153 | ||
| 154 | crypto | |
| 155 | encryption/decryption | |
| 156 | signature verification | |
| 157 | key usage | |
| 158 | nonce/IV rules | |
| 159 | hashing or key derivation | |
| 160 | ||
| 161 | input-validation | |
| 162 | validation of untrusted inputs | |
| 163 | canonicalization | |
| 164 | malformed input rejection | |
| 165 | path normalization | |
| 166 | ||
| 167 | injection | |
| 168 | protection against injection attacks | |
| 169 | SQL/NoSQL injection | |
| 170 | command injection | |
| 171 | template injection | |
| 172 | deserialization vulnerabilities | |
| 173 | ||
| 174 | data-protection | |
| 175 | encryption of sensitive data | |
| 176 | secret handling | |
| 177 | PII protection | |
| 178 | secure storage | |
| 179 | ||
| 180 | logging | |
| 181 | security event logging | |
| 182 | audit events | |
| 183 | absence of secrets in logs | |
| 184 | ||
| 185 | error-handling | |
| 186 | safe error messages | |
| 187 | no information leakage | |
| 188 | safe fallback behavior | |
| 189 | ||
| 190 | ||
| 191 | OWASP Tag | |
| 192 | --------- | |
| 193 | ||
| 194 | The `owasp` tag indicates that the test validates protection against a vulnerability | |
| 195 | category commonly described in OWASP guidance such as: | |
| 196 | ||
| 197 | • injection | |
| 198 | • broken authentication | |
| 199 | • broken access control | |
| 200 | • security misconfiguration | |
| 201 | • sensitive data exposure | |
| 202 | • insecure deserialization | |
| 203 | • cross-site scripting | |
| 204 | ||
| 205 | The `owasp` tag should only be used when the test clearly targets a known | |
| 206 | OWASP vulnerability category. | |
| 207 | ||
| 208 | Prefer combining `owasp` with a more precise taxonomy tag. | |
| 209 | ||
| 210 | ||
| 211 | Tag Selection Rules | |
| 212 | ------------------- | |
| 213 | ||
| 214 | 1. If a test validates a security property → include `security`. | |
| 215 | 2. Add 1–3 additional category tags when applicable. | |
| 216 | 3. Prefer the most specific tag. | |
| 217 | 4. Do not assign tags when security relevance is unclear. | |
| 218 | 5. Never invent new tags. | |
| 219 | ||
| 220 | ||
| 221 | Display Name Format | |
| 222 | ------------------- | |
| 223 | ||
| 224 | SECURITY: <security property> - <scenario> | |
| 225 | ||
| 226 | Examples: | |
| 227 | ||
| 228 | SECURITY: access control - deny non-owner account access | |
| 229 | SECURITY: crypto - reject reused nonce in AEAD | |
| 230 | SECURITY: input validation - reject path traversal sequences | |
| 231 | """; | |
| 232 | } | |
| 233 | } | |
Mutations | ||
| 79 |
1.1 |