pwn 0.5.795 → 0.5.796

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (369) hide show
  1. checksums.yaml +4 -4
  2. data/Gemfile +1 -1
  3. data/bin/pwn_autoinc_version +4 -5
  4. data/bin/pwn_web_cache_deception +40 -29
  5. data/documentation/AWS.md +15 -5
  6. data/documentation/Diagrams.md +1 -1
  7. data/documentation/Home.md +2 -2
  8. data/documentation/How-PWN-Works.md +1 -1
  9. data/documentation/What-is-PWN.md +1 -1
  10. data/documentation/diagrams/aws-cloud-security.svg +1 -1
  11. data/documentation/diagrams/dot/aws-cloud-security.dot +1 -1
  12. data/documentation/diagrams/dot/network-infra-testing.dot +1 -1
  13. data/documentation/diagrams/network-infra-testing.svg +1 -1
  14. data/etc/default_cron/session_cleanup.rb +84 -51
  15. data/lib/pwn/ai/agent/assembly.rb +0 -2
  16. data/lib/pwn/ai/agent/btc.rb +0 -2
  17. data/lib/pwn/ai/agent/burp_suite.rb +0 -2
  18. data/lib/pwn/ai/agent/curriculum.rb +2 -2
  19. data/lib/pwn/ai/agent/dispatch.rb +3 -0
  20. data/lib/pwn/ai/agent/engagement.rb +96 -41
  21. data/lib/pwn/ai/agent/extrospection/osint/bridges.rb +2 -0
  22. data/lib/pwn/ai/agent/extrospection.rb +50 -13
  23. data/lib/pwn/ai/agent/gqrx.rb +0 -2
  24. data/lib/pwn/ai/agent/hacker_one.rb +0 -2
  25. data/lib/pwn/ai/agent/learning.rb +45 -13
  26. data/lib/pwn/ai/agent/loop.rb +12 -11
  27. data/lib/pwn/ai/agent/metrics.rb +1 -1
  28. data/lib/pwn/ai/agent/mission.rb +6 -3
  29. data/lib/pwn/ai/agent/policy.rb +1 -1
  30. data/lib/pwn/ai/agent/policy_evaluation.rb +2 -2
  31. data/lib/pwn/ai/agent/reflect.rb +1 -0
  32. data/lib/pwn/ai/agent/request_runtime.rb +1 -1
  33. data/lib/pwn/ai/agent/reward.rb +155 -53
  34. data/lib/pwn/ai/agent/sast.rb +0 -2
  35. data/lib/pwn/ai/agent/task_dag.rb +5 -3
  36. data/lib/pwn/ai/agent/task_summarizer.rb +21 -11
  37. data/lib/pwn/ai/agent/tool_guard.rb +19 -0
  38. data/lib/pwn/ai/agent/tools/ruby_eval.rb +1 -1
  39. data/lib/pwn/ai/agent/tools/verify.rb +7 -3
  40. data/lib/pwn/ai/agent/transparent_browser.rb +0 -2
  41. data/lib/pwn/ai/agent/turn_finalizer.rb +3 -0
  42. data/lib/pwn/ai/anthropic.rb +3 -3
  43. data/lib/pwn/ai/context.rb +2 -1
  44. data/lib/pwn/ai/gemini.rb +7 -5
  45. data/lib/pwn/ai/grok.rb +37 -18
  46. data/lib/pwn/ai/mcp.rb +17 -5
  47. data/lib/pwn/ai/ollama.rb +76 -35
  48. data/lib/pwn/ai/open_ai.rb +7 -19
  49. data/lib/pwn/ai/open_web_ui.rb +76 -40
  50. data/lib/pwn/aws/cloud_search_domain.rb +6 -0
  51. data/lib/pwn/aws/iam.rb +2 -2
  52. data/lib/pwn/aws/workspaces.rb +2 -2
  53. data/lib/pwn/aws.rb +0 -5
  54. data/lib/pwn/bounty/lifecycle_authz_replay.rb +1 -1
  55. data/lib/pwn/config.rb +17 -10
  56. data/lib/pwn/corpus.rb +33 -5
  57. data/lib/pwn/cron.rb +60 -17
  58. data/lib/pwn/driver.rb +2 -3
  59. data/lib/pwn/engagement.rb +28 -10
  60. data/lib/pwn/ffi/capstone.rb +3 -2
  61. data/lib/pwn/ffi/fftw.rb +8 -6
  62. data/lib/pwn/ffi/keystone.rb +3 -2
  63. data/lib/pwn/ffi/soapy_sdr.rb +10 -6
  64. data/lib/pwn/json_text.rb +25 -13
  65. data/lib/pwn/memory_index.rb +6 -4
  66. data/lib/pwn/module_skills.rb +3 -1
  67. data/lib/pwn/plugins/aflplusplus.rb +1 -0
  68. data/lib/pwn/plugins/ai_sandbox.rb +42 -10
  69. data/lib/pwn/plugins/android.rb +429 -466
  70. data/lib/pwn/plugins/artifact_registry.rb +12 -2
  71. data/lib/pwn/plugins/assembly.rb +9 -42
  72. data/lib/pwn/plugins/authentication_helper.rb +4 -3
  73. data/lib/pwn/plugins/baresip.rb +70 -28
  74. data/lib/pwn/plugins/binary_analysis.rb +14 -5
  75. data/lib/pwn/plugins/binary_parser.rb +42 -7
  76. data/lib/pwn/plugins/black_duck_binary_analysis.rb +3 -3
  77. data/lib/pwn/plugins/burp_suite.rb +15 -12
  78. data/lib/pwn/plugins/bus_pirate.rb +3 -7
  79. data/lib/pwn/plugins/char.rb +9 -3
  80. data/lib/pwn/plugins/credit_card.rb +7 -0
  81. data/lib/pwn/plugins/dao_ldap.rb +2 -2
  82. data/lib/pwn/plugins/dao_mongo.rb +1 -1
  83. data/lib/pwn/plugins/debugger.rb +1 -1
  84. data/lib/pwn/plugins/defect_dojo.rb +7 -3
  85. data/lib/pwn/plugins/detect_os.rb +1 -1
  86. data/lib/pwn/plugins/detonate.rb +9 -19
  87. data/lib/pwn/plugins/exploit_db.rb +1 -1
  88. data/lib/pwn/plugins/file_fu.rb +4 -1
  89. data/lib/pwn/plugins/findings.rb +76 -2
  90. data/lib/pwn/plugins/fuzz.rb +12 -3
  91. data/lib/pwn/plugins/gdb.rb +7 -2
  92. data/lib/pwn/plugins/gdbmi.rb +10 -6
  93. data/lib/pwn/plugins/git.rb +18 -13
  94. data/lib/pwn/plugins/github.rb +11 -6
  95. data/lib/pwn/plugins/google_workspace.rb +11 -3
  96. data/lib/pwn/plugins/hacker_one.rb +5 -1
  97. data/lib/pwn/plugins/httpx.rb +24 -8
  98. data/lib/pwn/plugins/ip_info.rb +11 -8
  99. data/lib/pwn/plugins/irc.rb +73 -45
  100. data/lib/pwn/plugins/jenkins.rb +2 -2
  101. data/lib/pwn/plugins/jira_data_center.rb +13 -6
  102. data/lib/pwn/plugins/jobs.rb +62 -4
  103. data/lib/pwn/plugins/log.rb +1 -1
  104. data/lib/pwn/plugins/monkey_patch.rb +1 -1
  105. data/lib/pwn/plugins/msr206.rb +157 -182
  106. data/lib/pwn/plugins/nessus_cloud.rb +2 -2
  107. data/lib/pwn/plugins/nexpose_vuln_scan.rb +44 -33
  108. data/lib/pwn/plugins/nuclei.rb +24 -8
  109. data/lib/pwn/plugins/ocr.rb +3 -1
  110. data/lib/pwn/plugins/open_api.rb +0 -4
  111. data/lib/pwn/plugins/openvas.rb +31 -43
  112. data/lib/pwn/plugins/packet.rb +40 -25
  113. data/lib/pwn/plugins/pony.rb +2 -2
  114. data/lib/pwn/plugins/preflight_checker.rb +1 -1
  115. data/lib/pwn/plugins/process_tube.rb +17 -2
  116. data/lib/pwn/plugins/ps.rb +6 -1
  117. data/lib/pwn/plugins/repl/ai.rb +1 -0
  118. data/lib/pwn/plugins/repl/ai_console.rb +12 -5
  119. data/lib/pwn/plugins/repl/mesh.rb +10 -2
  120. data/lib/pwn/plugins/sbom.rb +59 -27
  121. data/lib/pwn/plugins/scannable_codes.rb +3 -1
  122. data/lib/pwn/plugins/serial.rb +102 -44
  123. data/lib/pwn/plugins/shodan.rb +4 -10
  124. data/lib/pwn/plugins/sock.rb +16 -11
  125. data/lib/pwn/plugins/ssn.rb +3 -1
  126. data/lib/pwn/plugins/transparent_browser.rb +43 -3
  127. data/lib/pwn/plugins/vault.rb +92 -30
  128. data/lib/pwn/plugins/vin.rb +2 -1
  129. data/lib/pwn/plugins/vsphere.rb +2 -1
  130. data/lib/pwn/plugins/xxd.rb +3 -5
  131. data/lib/pwn/plugins/yubikey/security_domain/scp.rb +4 -0
  132. data/lib/pwn/plugins/yubikey/transport.rb +42 -11
  133. data/lib/pwn/plugins/yubikey.rb +21 -6
  134. data/lib/pwn/plugins/zaproxy.rb +3 -4
  135. data/lib/pwn/reports/ai_red_team.rb +4 -4
  136. data/lib/pwn/reports/pdf.rb +9 -5
  137. data/lib/pwn/reports/sast.rb +4 -4
  138. data/lib/pwn/sast/keystore.rb +2 -2
  139. data/lib/pwn/sast/logger.rb +1 -1
  140. data/lib/pwn/sast/php_type_juggling.rb +3 -6
  141. data/lib/pwn/sast/pom_version.rb +12 -6
  142. data/lib/pwn/sast/port.rb +1 -1
  143. data/lib/pwn/sast/sql.rb +3 -7
  144. data/lib/pwn/sast/test_case_engine.rb +62 -16
  145. data/lib/pwn/sast/type_script_type_juggling.rb +3 -6
  146. data/lib/pwn/sdr/decoder/base.rb +8 -0
  147. data/lib/pwn/sdr/decoder/lora.rb +1 -1
  148. data/lib/pwn/sdr/flipper_zero.rb +4 -7
  149. data/lib/pwn/sdr/gqrx.rb +35 -16
  150. data/lib/pwn/sdr/son_micro_rfid.rb +44 -26
  151. data/lib/pwn/sdr.rb +1 -1
  152. data/lib/pwn/session_trace.rb +7 -2
  153. data/lib/pwn/sessions.rb +307 -39
  154. data/lib/pwn/setup.rb +4 -3
  155. data/lib/pwn/version.rb +1 -1
  156. data/lib/pwn/www/app_cobalt_io.rb +14 -2
  157. data/lib/pwn/www/bing.rb +13 -1
  158. data/lib/pwn/www/bug_crowd.rb +14 -2
  159. data/lib/pwn/www/checkip.rb +13 -1
  160. data/lib/pwn/www/coinbase_pro.rb +14 -2
  161. data/lib/pwn/www/duckduckgo.rb +13 -1
  162. data/lib/pwn/www/facebook.rb +14 -2
  163. data/lib/pwn/www/github.rb +14 -2
  164. data/lib/pwn/www/google.rb +13 -1
  165. data/lib/pwn/www/hacker_one.rb +117 -33
  166. data/lib/pwn/www/linkedin.rb +14 -2
  167. data/lib/pwn/www/pandora.rb +14 -2
  168. data/lib/pwn/www/pastebin.rb +13 -1
  169. data/lib/pwn/www/paypal.rb +16 -4
  170. data/lib/pwn/www/synack.rb +14 -2
  171. data/lib/pwn/www/torch.rb +17 -3
  172. data/lib/pwn/www/trading_view.rb +14 -2
  173. data/lib/pwn/www/twitter.rb +14 -2
  174. data/lib/pwn/www/uber.rb +14 -2
  175. data/lib/pwn/www/upwork.rb +14 -2
  176. data/lib/pwn/www/wayback_machine.rb +16 -2
  177. data/lib/pwn/www/youtube.rb +13 -1
  178. data/pwn.gemspec +4 -0
  179. data/spec/integration/authentication_helper_oath_spec.rb +1 -1
  180. data/spec/integration/autoinc_lifecycle_supplement_spec.rb +47 -0
  181. data/spec/integration/gemspec_deleted_files_spec.rb +43 -0
  182. data/spec/integration/spec_curriculum_isolation_spec.rb +41 -0
  183. data/spec/integration/verify_acceptance_eui_regression_spec.rb +162 -0
  184. data/spec/integration/web_cache_lifecycle_closure_spec.rb +89 -0
  185. data/spec/integration/web_cache_lifecycle_supplement_spec.rb +70 -0
  186. data/spec/lib/pwn/ai/agent/analysis_errors_regression_spec.rb +33 -0
  187. data/spec/lib/pwn/ai/agent/curriculum_cooldown_regression_spec.rb +19 -0
  188. data/spec/lib/pwn/ai/agent/curriculum_spec.rb +2 -0
  189. data/spec/lib/pwn/ai/agent/engagement_boundary_regression_spec.rb +42 -0
  190. data/spec/lib/pwn/ai/agent/engagement_dispatch_review_spec.rb +82 -0
  191. data/spec/lib/pwn/ai/agent/engagement_roe_regression_spec.rb +25 -0
  192. data/spec/lib/pwn/ai/agent/engagement_scope_regression_spec.rb +37 -0
  193. data/spec/lib/pwn/ai/agent/extrospection/osint/bridges_spec.rb +4 -0
  194. data/spec/lib/pwn/ai/agent/extrospection_input_regression_spec.rb +31 -0
  195. data/spec/lib/pwn/ai/agent/extrospection_review_spec.rb +61 -0
  196. data/spec/lib/pwn/ai/agent/extrospection_serial_review_spec.rb +42 -0
  197. data/spec/lib/pwn/ai/agent/extrospection_spec.rb +9 -0
  198. data/spec/lib/pwn/ai/agent/extrospection_verification_regression_spec.rb +39 -0
  199. data/spec/lib/pwn/ai/agent/learning_last_answer_regression_spec.rb +32 -0
  200. data/spec/lib/pwn/ai/agent/learning_lifecycle_closure_spec.rb +84 -0
  201. data/spec/lib/pwn/ai/agent/learning_prune_regression_spec.rb +17 -0
  202. data/spec/lib/pwn/ai/agent/learning_spec.rb +2 -1
  203. data/spec/lib/pwn/ai/agent/loop_execution_audit_spec.rb +30 -0
  204. data/spec/lib/pwn/ai/agent/metrics_usage_regression_spec.rb +20 -0
  205. data/spec/lib/pwn/ai/agent/mission_execution_audit_spec.rb +27 -0
  206. data/spec/lib/pwn/ai/agent/mission_namespace_spec.rb +210 -18
  207. data/spec/lib/pwn/ai/agent/policy_confidence_regression_spec.rb +16 -0
  208. data/spec/lib/pwn/ai/agent/policy_evaluation_spec.rb +1 -1
  209. data/spec/lib/pwn/ai/agent/policy_evaluation_tmp_regression_spec.rb +36 -0
  210. data/spec/lib/pwn/ai/agent/reflect_model_regression_spec.rb +15 -0
  211. data/spec/lib/pwn/ai/agent/request_runtime_execution_audit_spec.rb +43 -0
  212. data/spec/lib/pwn/ai/agent/reward_binding_review_spec.rb +129 -0
  213. data/spec/lib/pwn/ai/agent/reward_dispatch_review_spec.rb +154 -0
  214. data/spec/lib/pwn/ai/agent/reward_envelope_regression_spec.rb +45 -0
  215. data/spec/lib/pwn/ai/agent/reward_turn_regression_spec.rb +40 -0
  216. data/spec/lib/pwn/ai/agent/swarm_spec.rb +2 -2
  217. data/spec/lib/pwn/ai/agent/task_dag_execution_audit_spec.rb +45 -0
  218. data/spec/lib/pwn/ai/agent/task_summarizer_failure_regression_spec.rb +26 -0
  219. data/spec/lib/pwn/ai/agent/tools/verify_execution_audit_spec.rb +33 -0
  220. data/spec/lib/pwn/ai/context_range_audit_spec.rb +24 -0
  221. data/spec/lib/pwn/ai/gemini_roundtrip_audit_spec.rb +33 -0
  222. data/spec/lib/pwn/ai/grok_reasoning_effort_spec.rb +43 -0
  223. data/spec/lib/pwn/ai/mcp_session_audit_spec.rb +36 -0
  224. data/spec/lib/pwn/ai/ollama_spec.rb +6 -11
  225. data/spec/lib/pwn/ai/open_web_ui_spec.rb +6 -11
  226. data/spec/lib/pwn/ai/provider_history_audit_spec.rb +44 -0
  227. data/spec/lib/pwn/ai/provider_operations_audit_spec.rb +53 -0
  228. data/spec/lib/pwn/ai/provider_review_history_stream_regression_spec.rb +143 -0
  229. data/spec/lib/pwn/ai/provider_review_tls_regression_spec.rb +62 -0
  230. data/spec/lib/pwn/ai/provider_stream_audit_spec.rb +49 -0
  231. data/spec/lib/pwn/aws/iam_decoding_regression_spec.rb +23 -0
  232. data/spec/lib/pwn/aws/services_runtime_regression_spec.rb +51 -0
  233. data/spec/lib/pwn/aws_spec.rb +6 -0
  234. data/spec/lib/pwn/banner_operational_audit_spec.rb +105 -0
  235. data/spec/lib/pwn/bounty/lifecycle_authz_replay_spec.rb +17 -0
  236. data/spec/lib/pwn/config_redaction_regression_spec.rb +24 -0
  237. data/spec/lib/pwn/config_skills_regression_spec.rb +38 -0
  238. data/spec/lib/pwn/corpus_core_review_regression_spec.rb +96 -0
  239. data/spec/lib/pwn/corpus_spec.rb +13 -0
  240. data/spec/lib/pwn/cron_core_review_regression_spec.rb +95 -0
  241. data/spec/lib/pwn/cron_execution_spec.rb +75 -0
  242. data/spec/lib/pwn/driver_core_review_regression_spec.rb +38 -0
  243. data/spec/lib/pwn/driver_spec.rb +13 -0
  244. data/spec/lib/pwn/engagement_core_review_regression_spec.rb +88 -0
  245. data/spec/lib/pwn/engagement_storage_regression_spec.rb +44 -0
  246. data/spec/lib/pwn/ffi/capstone_spec.rb +19 -0
  247. data/spec/lib/pwn/ffi/fftw_spec.rb +19 -0
  248. data/spec/lib/pwn/ffi/keystone_spec.rb +13 -0
  249. data/spec/lib/pwn/ffi/soapy_sdr_spec.rb +13 -0
  250. data/spec/lib/pwn/json_text_spec.rb +25 -0
  251. data/spec/lib/pwn/memory_index_spec.rb +23 -0
  252. data/spec/lib/pwn/module_skills_parsing_regression_spec.rb +34 -0
  253. data/spec/lib/pwn/plugins/ai_sandbox_boundary_review_spec.rb +57 -0
  254. data/spec/lib/pwn/plugins/ai_sandbox_spec.rb +10 -18
  255. data/spec/lib/pwn/plugins/android_review_fixes_spec.rb +121 -0
  256. data/spec/lib/pwn/plugins/android_transport_audit_spec.rb +64 -0
  257. data/spec/lib/pwn/plugins/api_import_namespace_audit_spec.rb +49 -0
  258. data/spec/lib/pwn/plugins/artifact_registry_namespace_audit_spec.rb +20 -0
  259. data/spec/lib/pwn/plugins/artifact_registry_review_fixes_spec.rb +63 -0
  260. data/spec/lib/pwn/plugins/assembly_namespace_audit_spec.rb +16 -0
  261. data/spec/lib/pwn/plugins/assembly_review_fixes_spec.rb +33 -0
  262. data/spec/lib/pwn/plugins/authentication_helper_spec.rb +14 -0
  263. data/spec/lib/pwn/plugins/baresip_review_fixes_spec.rb +127 -0
  264. data/spec/lib/pwn/plugins/baresip_transport_audit_spec.rb +59 -0
  265. data/spec/lib/pwn/plugins/binary_analysis_process_review_spec.rb +57 -0
  266. data/spec/lib/pwn/plugins/black_duck_namespace_audit_spec.rb +25 -0
  267. data/spec/lib/pwn/plugins/burp_suite_har_review_fixes_spec.rb +60 -0
  268. data/spec/lib/pwn/plugins/burp_suite_import_review_fixes_spec.rb +62 -0
  269. data/spec/lib/pwn/plugins/burp_suite_namespace_audit_spec.rb +37 -0
  270. data/spec/lib/pwn/plugins/bus_pirate_transport_audit_spec.rb +19 -0
  271. data/spec/lib/pwn/plugins/file_fu_namespace_audit_spec.rb +17 -0
  272. data/spec/lib/pwn/plugins/findings_structured_spec.rb +1 -1
  273. data/spec/lib/pwn/plugins/gdbmi_namespace_audit_spec.rb +28 -0
  274. data/spec/lib/pwn/plugins/gdbmi_review_fixes_spec.rb +73 -0
  275. data/spec/lib/pwn/plugins/github_argv_review_spec.rb +49 -0
  276. data/spec/lib/pwn/plugins/google_workspace_provider_review_oauth_spec.rb +80 -0
  277. data/spec/lib/pwn/plugins/irc_review_fixes_spec.rb +92 -0
  278. data/spec/lib/pwn/plugins/irc_transport_audit_spec.rb +37 -0
  279. data/spec/lib/pwn/plugins/jenkins_namespace_audit_spec.rb +12 -0
  280. data/spec/lib/pwn/plugins/jira_data_center_namespace_audit_spec.rb +44 -0
  281. data/spec/lib/pwn/plugins/jira_data_center_review_fixes_spec.rb +81 -0
  282. data/spec/lib/pwn/plugins/jsonl_elf_review_fixes_spec.rb +59 -0
  283. data/spec/lib/pwn/plugins/log_spec.rb +36 -20
  284. data/spec/lib/pwn/plugins/monkey_patch_spec.rb +16 -0
  285. data/spec/lib/pwn/plugins/msr206_readback_review_fixes_spec.rb +119 -0
  286. data/spec/lib/pwn/plugins/msr206_review_fixes_spec.rb +50 -0
  287. data/spec/lib/pwn/plugins/msr206_transport_audit_spec.rb +64 -0
  288. data/spec/lib/pwn/plugins/nessus_cloud_namespace_audit_spec.rb +14 -0
  289. data/spec/lib/pwn/plugins/nexpose_review_fixes_spec.rb +55 -0
  290. data/spec/lib/pwn/plugins/open_api_namespace_audit_spec.rb +22 -0
  291. data/spec/lib/pwn/plugins/packet_transport_audit_spec.rb +47 -0
  292. data/spec/lib/pwn/plugins/plugins_b_operational_regression_spec.rb +600 -0
  293. data/spec/lib/pwn/plugins/pony_namespace_audit_spec.rb +17 -0
  294. data/spec/lib/pwn/plugins/pony_review_fixes_spec.rb +33 -0
  295. data/spec/lib/pwn/plugins/process_tube_namespace_audit_spec.rb +19 -0
  296. data/spec/lib/pwn/plugins/ps_namespace_audit_spec.rb +12 -0
  297. data/spec/lib/pwn/plugins/repl/ai_console_spec.rb +111 -0
  298. data/spec/lib/pwn/plugins/repl/ai_spec.rb +38 -0
  299. data/spec/lib/pwn/plugins/repl/mesh_spec.rb +52 -0
  300. data/spec/lib/pwn/plugins/sandbox_spec.rb +9 -7
  301. data/spec/lib/pwn/plugins/sbom_review_fixes_spec.rb +43 -0
  302. data/spec/lib/pwn/plugins/scannable_codes_namespace_audit_spec.rb +15 -0
  303. data/spec/lib/pwn/plugins/serial_review_fixes_spec.rb +76 -0
  304. data/spec/lib/pwn/plugins/serial_transport_audit_spec.rb +102 -0
  305. data/spec/lib/pwn/plugins/shodan_namespace_audit_spec.rb +10 -0
  306. data/spec/lib/pwn/plugins/sock_transport_audit_spec.rb +55 -0
  307. data/spec/lib/pwn/plugins/transparent_browser_navigation_review_spec.rb +21 -0
  308. data/spec/lib/pwn/plugins/vault_master_key_eui_regression_spec.rb +242 -0
  309. data/spec/lib/pwn/plugins/vault_spec.rb +58 -0
  310. data/spec/lib/pwn/plugins/vault_store_safety_spec.rb +1 -1
  311. data/spec/lib/pwn/plugins/xxd_namespace_audit_spec.rb +34 -0
  312. data/spec/lib/pwn/plugins/yubikey_hid_cleanup_review_fixes_spec.rb +52 -0
  313. data/spec/lib/pwn/plugins/yubikey_lifecycle_audit_spec.rb +65 -0
  314. data/spec/lib/pwn/plugins/yubikey_options_audit_spec.rb +30 -0
  315. data/spec/lib/pwn/plugins/yubikey_selector_audit_spec.rb +26 -0
  316. data/spec/lib/pwn/plugins/yubikey_spec.rb +1 -1
  317. data/spec/lib/pwn/plugins/zaproxy_namespace_audit_spec.rb +30 -0
  318. data/spec/lib/pwn/reports/empty_report_audit_spec.rb +18 -0
  319. data/spec/lib/pwn/reports/pdf_spec.rb +25 -0
  320. data/spec/lib/pwn/reports/provider_review_encoding_regression_spec.rb +45 -0
  321. data/spec/lib/pwn/sast/engine_runtime_regression_spec.rb +91 -0
  322. data/spec/lib/pwn/sast/filter_edge_regression_spec.rb +39 -0
  323. data/spec/lib/pwn/sast/formatter_review_fixes_spec.rb +178 -0
  324. data/spec/lib/pwn/sast/path_review_fixes_spec.rb +60 -0
  325. data/spec/lib/pwn/sast/pom_runtime_regression_spec.rb +37 -0
  326. data/spec/lib/pwn/sast/rules_runtime_regression_spec.rb +78 -0
  327. data/spec/lib/pwn/sast/sql_review_fixes_spec.rb +38 -0
  328. data/spec/lib/pwn/sdr/decoder/base_iq_spec.rb +21 -18
  329. data/spec/lib/pwn/sdr/decoder/base_review_fixes_spec.rb +49 -0
  330. data/spec/lib/pwn/sdr/decoder/base_validation_spec.rb +78 -0
  331. data/spec/lib/pwn/sdr/decoder/lora_spec.rb +9 -0
  332. data/spec/lib/pwn/sdr/flipper_zero_review_fixes_spec.rb +34 -0
  333. data/spec/lib/pwn/sdr/gqrx_operational_spec.rb +83 -0
  334. data/spec/lib/pwn/sdr/gqrx_review_fixes_spec.rb +22 -0
  335. data/spec/lib/pwn/sdr/son_micro_rfid_review_fixes_spec.rb +113 -0
  336. data/spec/lib/pwn/sdr/son_micro_rfid_serial_review_fixes_spec.rb +76 -0
  337. data/spec/lib/pwn/sdr/son_micro_rfid_spec.rb +39 -0
  338. data/spec/lib/pwn/sdr_spec.rb +5 -0
  339. data/spec/lib/pwn/session_cleanup_default_spec.rb +1 -1
  340. data/spec/lib/pwn/session_cleanup_lifecycle_supplement_spec.rb +39 -0
  341. data/spec/lib/pwn/session_cleanup_payload_spec.rb +3 -2
  342. data/spec/lib/pwn/session_trace_lifecycle_supplement_spec.rb +39 -0
  343. data/spec/lib/pwn/sessions_archive_closure_spec.rb +126 -0
  344. data/spec/lib/pwn/sessions_lifecycle_closure_spec.rb +373 -0
  345. data/spec/lib/pwn/sessions_lifecycle_supplement_spec.rb +204 -0
  346. data/spec/lib/pwn/setup_operational_regression_spec.rb +36 -0
  347. data/spec/lib/pwn/www/browser_review_fixes_spec.rb +70 -0
  348. data/spec/lib/pwn/www/browser_runtime_regression_spec.rb +100 -0
  349. data/spec/lib/pwn/www/hacker_one_runtime_regression_spec.rb +146 -0
  350. data/spec/lib/pwn/www/pagination_review_fixes_spec.rb +84 -0
  351. data/spec/support/namespace_controller_observability_fixture.rb +141 -0
  352. data/spec/support/sandbox.rb +8 -5
  353. data/third_party/pwn_rdoc.jsonl +73 -31
  354. metadata +137 -18
  355. data/etc/default_skills/pwn/aws/elastic_transcoder/SKILL.md +0 -48
  356. data/etc/default_skills/pwn/aws/lambda_preview/SKILL.md +0 -48
  357. data/etc/default_skills/pwn/aws/ops_works/SKILL.md +0 -48
  358. data/etc/default_skills/pwn/aws/ops_works_cm/SKILL.md +0 -48
  359. data/etc/default_skills/pwn/aws/sms/SKILL.md +0 -48
  360. data/lib/pwn/aws/elastic_transcoder.rb +0 -95
  361. data/lib/pwn/aws/lambda_preview.rb +0 -95
  362. data/lib/pwn/aws/ops_works.rb +0 -95
  363. data/lib/pwn/aws/ops_works_cm.rb +0 -95
  364. data/lib/pwn/aws/sms.rb +0 -95
  365. data/spec/lib/pwn/aws/elastic_transcoder_spec.rb +0 -15
  366. data/spec/lib/pwn/aws/lambda_preview_spec.rb +0 -15
  367. data/spec/lib/pwn/aws/ops_works_cm_spec.rb +0 -15
  368. data/spec/lib/pwn/aws/ops_works_spec.rb +0 -15
  369. data/spec/lib/pwn/aws/sms_spec.rb +0 -15
@@ -0,0 +1,154 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+ require 'pwn/ai/agent/tools/ruby_eval'
5
+ require 'pwn/ai/agent/tools/shell'
6
+
7
+ RSpec.describe 'AI learning dispatch result review' do
8
+ include_context 'pwn tmp sandbox'
9
+
10
+ before do
11
+ allow(Dir).to receive(:home).and_return(@tmp)
12
+ @agent_cfg[:reward_llm] = false
13
+ PWN::Env[:ai][:sandbox] = 'off'
14
+ @saved_runtime = Thread.current[:pwn_request_runtime]
15
+ Thread.current[:pwn_request_runtime] = nil
16
+ stub_const('PWN::AI::Agent::Engagement::ROOT', File.join(@tmp, 'engagements'))
17
+ stub_const('PWN::AI::Agent::Engagement::ACTIVE_FILE', File.join(@tmp, 'engagements', 'active'))
18
+ end
19
+
20
+ after do
21
+ Thread.current[:pwn_request_runtime] = @saved_runtime
22
+ end
23
+
24
+ def dispatch(name, args)
25
+ PWN::AI::Agent::Dispatch.call(tool_call: { id: 'outcome-fixture', function: { name: name, arguments: JSON.generate(args) } })
26
+ end
27
+
28
+ ["{success: false, error: 'File.write failed in the fixture'}", "{'ok' => false}"].each do |code|
29
+ it "separates completed evaluation from application failure for #{code}" do
30
+ raw = dispatch('pwn_eval', code: code, timeout: 3)
31
+ result = JSON.parse(raw)
32
+ expect(result.dig('result', 'value')).to include('false')
33
+ expect(PWN::AI::Agent::Reward.semantic_ok(name: 'pwn_eval', raw: raw)).to include(semantic_ok: false)
34
+ expect(result.dig('result', 'execution_success')).to be(true)
35
+ expect(result.dig('result', 'application_status')).to eq('failed')
36
+ state = PWN::AI::Agent::TaskSummarizer.fresh(request: 'Write the required report')
37
+ state[:plan] = ['Write the required report', 'Present the report']
38
+ PWN::AI::Agent::TaskSummarizer.record!(state: state, name: 'pwn_eval', args: { code: code }, result: raw)
39
+ expect(state[:plan_idx]).to eq(0)
40
+ expect(PWN::AI::Agent::Reward.prm(request: state[:request], trace: ["pwn_eval → #{raw}"]).first[:reward]).to eq(-1)
41
+ end
42
+ end
43
+
44
+ it 'does not let a provider override an observed application failure with positive credit' do
45
+ @agent_cfg[:reward_llm] = true
46
+ allow(PWN::AI::Ollama).to receive(:chat).and_return('1')
47
+ raw = dispatch('pwn_eval', code: '{success: false}', timeout: 3)
48
+ expect(PWN::AI::Agent::Reward.prm(request: 'original request', trace: ["pwn_eval → #{raw}"]).first[:reward]).to be <= 0
49
+ end
50
+
51
+ it 'keeps a Loop-capped successful observation unknown rather than inventing a negative reward' do
52
+ PWN::Env[:ai][:active] = :openai
53
+ raw = dispatch('pwn_eval', code: "'a' * 1500", timeout: 3)
54
+ expect(PWN::AI::Agent::Reward.semantic_ok(name: 'pwn_eval', raw: raw)[:semantic_ok]).to be(true)
55
+ conditioned = PWN::AI::Agent::Result.condition(content: raw, entry: PWN::AI::Agent::Registry.lookup(name: 'pwn_eval'))
56
+ cap = PWN::AI::Agent::Loop.send(:session_tool_budget, name: 'pwn_eval', result: conditioned)
57
+ sid = PWN::Sessions.create(title: 'capped result fixture')[:id]
58
+ PWN::Sessions.append(session_id: sid, role: 'user', content: 'original request')
59
+ PWN::Sessions.append(session_id: sid, role: 'tool', content: "pwn_eval → #{conditioned[0, cap]}")
60
+ rows = PWN::AI::Agent::Reward.prm(request: 'original request', session_id: sid)
61
+ expect(rows.first[:reward]).to be_nil
62
+ expect(PWN::Sessions.load(session_id: sid).last[:step_reward]).to be_nil
63
+ expect(PWN::AI::Agent::Metrics.prm_n(name: 'pwn_eval')).to eq(0)
64
+ end
65
+
66
+ it 'keeps malformed and untyped status evidence unknown even if a judge returns a score' do
67
+ @agent_cfg[:reward_llm] = true
68
+ allow(PWN::AI::Ollama).to receive(:chat).and_return('1')
69
+ ['shell → {"success":true', 'shell → timeout success:false', 'shell → {"result":{"stdout":"success"}}'].each do |trace|
70
+ expect(PWN::AI::Agent::Reward.semantic_ok(name: 'shell', raw: trace)).to include(semantic_ok: nil, shape: :unknown)
71
+ expect(PWN::AI::Agent::Reward.prm(request: 'original request', trace: [trace]).first[:reward]).to be_nil
72
+ end
73
+ end
74
+
75
+ it 'does not persist unknown capped observations as Loop mistakes or failure metrics' do
76
+ raw = dispatch('pwn_eval', code: "'a' * 1500", timeout: 3)
77
+ verdict = PWN::AI::Agent::Loop.send(:record_metrics, name: 'pwn_eval', raw: raw[0, 64], args: { code: 'inert fixture' })
78
+ expect(verdict).to include(ok: nil, mistake: nil)
79
+ expect(PWN::AI::Agent::Metrics.summary).to be_empty
80
+ expect(File).not_to exist(PWN::AI::Agent::Mistakes::MISTAKES_FILE)
81
+ end
82
+
83
+ it 'classifies an actual informational grep miss without a false timeout fingerprint' do
84
+ args = { command: 'grep absent /dev/null', timeout: 3 }
85
+ raw = dispatch('shell', args)
86
+ expect(JSON.parse(raw).dig('result', 'exit')).to eq(1)
87
+ expect(PWN::AI::Agent::Reward.semantic_ok(name: 'shell', raw: raw, args: args)).to include(ok: false, semantic_ok: true, benign: true, shape: nil)
88
+ end
89
+
90
+ it 'preserves informational shell classification in the retained PRM trace without argv' do
91
+ raw = dispatch('shell', command: 'grep absent /dev/null', timeout: 3)
92
+ sid = PWN::Sessions.create(title: 'informational observation')[:id]
93
+ PWN::Sessions.append(session_id: sid, role: 'user', content: 'Find a matching record')
94
+ PWN::Sessions.append(session_id: sid, role: 'tool', content: "shell → #{raw}")
95
+ steps = PWN::AI::Agent::Reward.prm(request: 'Find a matching record', session_id: sid)
96
+ expect(steps.first[:reward]).to eq(0)
97
+ expect(PWN::Sessions.load(session_id: sid).last[:step_reward]).to eq(0)
98
+ end
99
+
100
+ it 'keeps real failures distinct from timeout configuration and informational exits' do
101
+ ['false', 'grep absent /dev/null | false', 'grep absent /nonexistent-fixture'].each do |command|
102
+ raw = dispatch('shell', command: command, timeout: 3)
103
+ verdict = PWN::AI::Agent::Reward.semantic_ok(name: 'shell', raw: raw, args: { command: command })
104
+ expect(verdict).to include(semantic_ok: false, benign: false)
105
+ expect(verdict[:shape]).not_to eq(:timeout)
106
+ end
107
+ end
108
+
109
+ it 'does not credit failed write text to a later real successful no-op' do
110
+ state = PWN::AI::Agent::TaskSummarizer.fresh(request: 'Write the required report')
111
+ state[:plan] = ['Write the required report', 'Present the report']
112
+ ["{success: false, error: 'File.write could not create report'}", '1 + 1'].each do |code|
113
+ raw = dispatch('pwn_eval', code: code, timeout: 3)
114
+ PWN::AI::Agent::TaskSummarizer.record!(state: state, name: 'pwn_eval', args: { code: code }, result: raw)
115
+ expect(state[:plan_idx]).to eq(0)
116
+ end
117
+ expect(state[:events].map { |row| row[:ok] }).to eq([false, true])
118
+ expect(state[:request]).to eq('Write the required report')
119
+ end
120
+
121
+ it 'does not treat mutation words in successful no-op output as an observed write effect' do
122
+ state = PWN::AI::Agent::TaskSummarizer.fresh(request: 'Write the required report')
123
+ state[:plan] = ['Write the required report', 'Present the report']
124
+ code = "'wrote the report'"
125
+ raw = dispatch('pwn_eval', code: code, timeout: 3)
126
+ PWN::AI::Agent::TaskSummarizer.record!(state: state, name: 'pwn_eval', args: { code: code }, result: raw)
127
+ expect(state[:plan_idx]).to eq(0)
128
+ expect(PWN::AI::Agent::TaskSummarizer.unfinished_tasks(state: state).map { |row| row[:idx] }).to include(0)
129
+ end
130
+
131
+ it 'preserves typed application failure across the sandbox boundary', :sandbox_outcome_boundary do
132
+ PWN::Env[:ai][:sandbox] = 'permissive'
133
+ expect(PWN::Plugins::AISandbox.mode).to eq(:permissive)
134
+ raw = dispatch('pwn_eval', code: '{success: false}', timeout: 3)
135
+ expect(JSON.parse(raw).dig('result', 'execution_success')).to be(true)
136
+ expect(JSON.parse(raw).dig('result', 'application_status')).to eq('failed')
137
+ expect(PWN::AI::Agent::Reward.semantic_ok(name: 'pwn_eval', raw: raw)[:semantic_ok]).to be(false)
138
+ end
139
+
140
+ it 'retains typed application and execution status even when a large result is paged' do
141
+ raw = dispatch('pwn_eval', code: "{success: false, details: 'x' * 10000}", timeout: 3)
142
+ parsed = JSON.parse(raw)
143
+ expect(parsed).to include('success' => false, 'execution_success' => true, 'application_status' => 'failed')
144
+ expect(parsed.dig('result', 'artifact')).not_to be_nil
145
+ end
146
+
147
+ it 'does not interpret stdout or inspected string values as application status' do
148
+ raw = dispatch('pwn_eval', code: "puts '{\"success\":false}'; '{success: false}'", timeout: 3)
149
+ result = JSON.parse(raw)
150
+ expect(result.dig('result', 'execution_success')).to be(true)
151
+ expect(result.dig('result', 'application_status')).to eq('unknown')
152
+ expect(PWN::AI::Agent::Reward.semantic_ok(name: 'pwn_eval', raw: raw)).to include(semantic_ok: true)
153
+ end
154
+ end
@@ -0,0 +1,45 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+
5
+ describe PWN::AI::Agent::Reward, 'structured result envelopes' do
6
+ it 'classifies pretty printed JSON the same as compact JSON' do
7
+ raw = JSON.pretty_generate(success: true, result: { exit: 0, stdout: 'fixture', stderr: '' })
8
+ expect(described_class.semantic_ok(name: 'shell', raw: raw)).to include(ok: true, semantic_ok: true, exit: 0)
9
+ end
10
+
11
+ it 'does not borrow a nested success flag for a failed dispatch' do
12
+ raw = JSON.generate(success: false, error: 'dispatch failed', result: { success: true })
13
+ expect(described_class.semantic_ok(name: 'shell', raw: raw)).to include(ok: false, semantic_ok: false)
14
+ end
15
+
16
+ it 'recognizes negative exits rather than treating them as absent' do
17
+ raw = JSON.generate(success: true, result: { exit: -1, stdout: '', stderr: '' })
18
+ expect(described_class.semantic_ok(name: 'shell', raw: raw)).to include(semantic_ok: false, exit: -1)
19
+ end
20
+
21
+ it 'does not classify observed timeout and invalid_payload text as an execution error' do
22
+ raw = JSON.generate(success: true, result: { exit: 0, stdout: 'documentation: timeout after 20s invalid_payload', stderr: '' })
23
+ expect(described_class.semantic_ok(name: 'shell', raw: "shell → #{raw}")).to include(semantic_ok: true)
24
+ end
25
+
26
+ it 'does not accept an explicitly failed nested result' do
27
+ raw = JSON.generate(success: true, result: { success: false, error: 'fixture failure' })
28
+ expect(described_class.semantic_ok(name: 'pwn_eval', raw: raw)).to include(ok: true, semantic_ok: false)
29
+ end
30
+
31
+ it 'does not treat a nested informational_exit flag as a dispatch envelope marker' do
32
+ raw = JSON.generate(success: false, result: { exit: 1, informational_exit: true, stdout: '', stderr: '' })
33
+ expect(described_class.semantic_ok(name: 'shell', raw: raw)).to include(semantic_ok: false, benign: false, shape: :nonzero_exit)
34
+ end
35
+
36
+ it 'does not let an envelope informational_exit launder a nested false result' do
37
+ raw = JSON.generate(success: false, informational_exit: true, result: { exit: 1, success: false, stdout: '', stderr: '' })
38
+ expect(described_class.semantic_ok(name: 'shell', raw: raw)).to include(semantic_ok: false, benign: false)
39
+ end
40
+
41
+ it 'keeps a dispatch envelope informational_exit benign without argv' do
42
+ raw = JSON.generate(success: false, informational_exit: true, result: { exit: 1, stdout: '', stderr: '' })
43
+ expect(described_class.semantic_ok(name: 'shell', raw: raw)).to include(ok: false, semantic_ok: true, benign: true, shape: nil)
44
+ end
45
+ end
@@ -0,0 +1,40 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+
5
+ describe PWN::AI::Agent::Reward, 'turn-scoped process rewards' do
6
+ include_context 'pwn tmp sandbox'
7
+
8
+ before do
9
+ allow(described_class).to receive(:llm_prm).and_return(nil)
10
+ @sid = PWN::Sessions.create(title: 'offline process reward fixture')[:id]
11
+ [
12
+ ['user', 'earlier request'],
13
+ ['tool', 'shell → {"success":true,"result":{"exit":0}}'],
14
+ ['assistant', 'earlier answer'],
15
+ ['user', 'current request'],
16
+ ['tool', 'pwn_eval → {"success":true,"result":{"exit":0}}'],
17
+ ['assistant', 'current answer']
18
+ ].each { |role, content| PWN::Sessions.append(session_id: @sid, role: role, content: content) }
19
+ end
20
+
21
+ it 'annotates the same latest-turn tools that were scored in the live session' do
22
+ described_class.prm(request: 'current request', session_id: @sid)
23
+ tools = PWN::Sessions.load(session_id: @sid).select { |row| row[:role] == 'tool' }
24
+ expect(tools.map { |row| row[:step_reward] }).to eq([nil, 1])
25
+ expect(PWN::AI::Agent::Metrics.prm_n(name: 'shell')).to eq(0)
26
+ expect(PWN::AI::Agent::Metrics.prm_n(name: 'pwn_eval')).to eq(1)
27
+ end
28
+
29
+ it 'annotates the same latest-turn tools in retained immutable evidence' do
30
+ learning = PWN::AI::Agent::Learning
31
+ learning.note_outcome(task: 'current request', success: true, session_id: @sid)
32
+ learning.migrate_evidence(dry_run: false)
33
+ File.unlink(File.join(PWN::Sessions::SESSIONS_DIR, "#{@sid}.jsonl"))
34
+ described_class.prm(request: 'current request', session_id: @sid)
35
+ tools = learning.session_evidence(session_id: @sid).select { |row| row[:role] == 'tool' }
36
+ expect(tools.map { |row| row[:step_reward] }).to eq([nil, 1])
37
+ expect(PWN::AI::Agent::Metrics.prm_n(name: 'shell')).to eq(0)
38
+ expect(PWN::AI::Agent::Metrics.prm_n(name: 'pwn_eval')).to eq(1)
39
+ end
40
+ end
@@ -242,9 +242,9 @@ describe PWN::AI::Agent::Swarm do
242
242
  end
243
243
 
244
244
  it 'maps several ports and returns a merged result set' do
245
+ expect(PWN::Plugins::Packet).to receive(:tcp_connect_scan).with(hosts: ['127.0.0.1'], ports: [1, 9, 22, 80]).and_return(results: [1, 9, 22, 80].map { |port| { host: '127.0.0.1', port: port, open: false } })
245
246
  rows = described_class.map_targets(targets: '127.0.0.1', ports: '1,9,22,80')
246
- expect(rows.length).to eq(4)
247
- expect(rows.map { |r| r[:port] }.uniq.length).to eq(4)
247
+ expect(rows).to eq([1, 9, 22, 80].map { |port| { host: '127.0.0.1', port: port, open: false } })
248
248
  end
249
249
 
250
250
  it 'allows only one live claim per unit' do
@@ -0,0 +1,45 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+
5
+ RSpec.describe PWN::AI::Agent::TaskDAG, 'execution failure reporting' do
6
+ include_context 'pwn tmp sandbox'
7
+
8
+ before do
9
+ allow(Dir).to receive(:home).and_return(@tmp)
10
+ end
11
+
12
+ it 'preserves Dispatch failure when a shell command exits nonzero without an error field' do
13
+ dag = described_class.plan(request: 'check command status', steps: [
14
+ { id: 'probe', tool: 'shell', args: { command: 'exit 7', timeout: 5 } }
15
+ ])
16
+ report = described_class.execute(dag: dag, root: @tmp, run_id: 'failed-exit')
17
+
18
+ expect(report[:results].first[:result]).to include(exit: 7)
19
+ expect(report[:results].first[:ok]).to be false
20
+ expect(report[:ok]).to be false
21
+ end
22
+
23
+ it 'halts at failed prerequisites and retries them rather than skipping them on resume' do
24
+ ready = File.join(@tmp, 'ready')
25
+ output = File.join(@tmp, 'dependent.txt')
26
+ dag = described_class.plan(request: 'wait for readiness then record it', steps: [
27
+ { id: 'ready', tool: 'shell', args: { command: "test -f #{Shellwords.escape(ready)}", timeout: 5 } },
28
+ { id: 'dependent', tool: 'shell', args: { command: "printf ready > #{Shellwords.escape(output)}", timeout: 5 }, dependencies: ['ready'] }
29
+ ])
30
+ report = described_class.execute(dag: dag, root: @tmp, run_id: 'retry-failed')
31
+
32
+ expect(report[:ok]).to be false
33
+ expect(File.exist?(output)).to be false
34
+ expect(report[:results].map { |row| row[:id] }).to eq(['ready'])
35
+ checkpoints = File.join(@tmp, 'retry-failed', 'checkpoints.jsonl')
36
+ expect(JSON.parse(File.read(checkpoints))).to include('id' => 'ready', 'status' => 'failed')
37
+
38
+ File.write(ready, 'ready')
39
+ resumed = described_class.resume(root: @tmp, run_id: 'retry-failed')
40
+ expect(resumed[:ok]).to be true
41
+ expect(resumed[:results].map { |row| row[:id] }).to eq(%w[ready dependent])
42
+ expect(File.read(output)).to eq('ready')
43
+ expect(File.readlines(checkpoints).map { |line| JSON.parse(line)['status'] }).to eq(%w[failed completed completed])
44
+ end
45
+ end
@@ -0,0 +1,26 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+
5
+ describe PWN::AI::Agent::TaskSummarizer, 'failed dispatch advancement' do
6
+ include_context 'pwn tmp sandbox'
7
+
8
+ it 'does not mark a failed write as completed because its error mentions File.write' do
9
+ state = described_class.fresh(request: 'write the required report')
10
+ state[:plan] = ['Write the report', 'Present the report']
11
+ result = JSON.generate(success: false, error: 'File.write could not create report')
12
+ described_class.record!(state: state, name: 'pwn_eval', args: { code: 'File.write(report, content)' }, result: result)
13
+ expect(state[:events].last[:ok]).to be(false)
14
+ expect(state[:plan_idx]).to eq(0)
15
+ end
16
+
17
+ it 'does not advance through the phase-handoff fallback after failed dispatches' do
18
+ state = described_class.fresh(request: 'inspect then patch the report')
19
+ state[:plan] = ['Inspect the report', 'Patch the report', 'Present the result']
20
+ 2.times do
21
+ described_class.record!(state: state, name: 'shell', args: { command: 'sed -i replacement report' }, result: '{"success":false,"error":"fixture failure"}')
22
+ end
23
+ expect(state[:plan_idx]).to eq(0)
24
+ expect(state[:last_prm_signal]).to eq(:hold_regress)
25
+ end
26
+ end
@@ -0,0 +1,33 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+
5
+ RSpec.describe 'verify file matching through Dispatch' do
6
+ include_context 'pwn tmp sandbox'
7
+
8
+ before do
9
+ allow(Dir).to receive(:home).and_return(@tmp)
10
+ PWN::AI::Agent::Registry.discover
11
+ end
12
+
13
+ it 'does not verify missing or directory paths as empty files' do
14
+ [File.join(@tmp, 'missing.txt'), @tmp].each do |path|
15
+ response = JSON.parse(PWN::AI::Agent::Dispatch.call(
16
+ tool_call: { function: { name: 'verify', arguments: JSON.generate(checks: [{ type: 'file_grep', path: path, pattern: '\\A\\z' }]) } }
17
+ ), symbolize_names: true)
18
+
19
+ expect(response.dig(:result, :passed)).to be(false), "incorrect verification for #{path}: #{response}"
20
+ expect(response.dig(:result, :results, 0)).to include(type: 'file_grep', passed: false, path: path)
21
+ end
22
+ end
23
+
24
+ it 'can verify an existing empty file with an empty-content expression' do
25
+ path = File.join(@tmp, 'empty.txt')
26
+ File.write(path, '')
27
+ response = JSON.parse(PWN::AI::Agent::Dispatch.call(
28
+ tool_call: { function: { name: 'verify', arguments: JSON.generate(checks: [{ type: 'file_grep', path: path, pattern: '\\A\\z' }]) } }
29
+ ), symbolize_names: true)
30
+
31
+ expect(response.dig(:result, :passed)).to be true
32
+ end
33
+ end
@@ -0,0 +1,24 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+ require 'tmpdir'
5
+ require 'digest'
6
+
7
+ RSpec.describe PWN::AI::Context, 'byte range regressions' do
8
+ { (0..0) => 'a', (1..3) => 'bcd', (1...3) => 'bc', (2...2) => '' }.each do |range, expected|
9
+ it "persists precisely the bytes selected by #{range.inspect}" do
10
+ Dir.mktmpdir('context-range-audit-') do |dir|
11
+ allow(Dir).to receive(:home).and_return(dir)
12
+ stub_const('PWN::Plugins::ArtifactRegistry::ROOT', File.join(dir, 'artifacts'))
13
+ path = File.join(dir, 'source.txt')
14
+ File.binwrite(path, 'abcdef')
15
+ result = described_class.attach_file(path: path, range: range, session_id: 'fixture-range')
16
+ expect(File.binread(result[:path])).to eq(expected)
17
+ expect(result[:chunks].join).to eq(expected)
18
+ expect(result[:size]).to eq(expected.bytesize)
19
+ expect(result[:sha256]).to eq(Digest::SHA256.hexdigest(expected))
20
+ expect(File.binread(path)).to eq('abcdef')
21
+ end
22
+ end
23
+ end
24
+ end
@@ -0,0 +1,33 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+ require 'rest-client'
5
+
6
+ RSpec.describe PWN::AI::Gemini, 'native content round trips' do
7
+ before do
8
+ stub_const('PWN::Env', { ai: { gemini: { model: 'gemini-fixture', key: 'fixture-key' } } })
9
+ allow(PWN::Plugins::TransparentBrowser).to receive(:open).and_return(browser: RestClient)
10
+ end
11
+
12
+ [false, true].each do |string_keys|
13
+ it "retains native part metadata and strips only local call IDs (string_keys=#{string_keys})" do
14
+ native = { functionCall: { name: 'fixture_tool', args: { value: 'source' } }, thoughtSignature: 'fixture-signature' }
15
+ requests = []
16
+ allow(RestClient::Request).to receive(:execute) do |args|
17
+ requests << JSON.parse(args.fetch(:payload))
18
+ body = { candidates: [{ content: { parts: [native] } }] }.to_json
19
+ instance_double(RestClient::Response, body: body)
20
+ end
21
+ first = described_class.chat_with_tools(messages: [{ role: 'user', content: 'fixture' }])
22
+ call_id = first.dig(:assistant_message, :tool_calls, 0, :id)
23
+ history = [first[:assistant_message], { role: 'tool', tool_call_id: call_id, content: '{"ok":true}' }]
24
+ history = JSON.parse(JSON.generate(history)) if string_keys
25
+ original = Marshal.dump(history)
26
+ described_class.chat_with_tools(messages: history)
27
+ contents = requests.last.fetch('contents')
28
+ expect(contents.first.fetch('parts')).to eq([JSON.parse(JSON.generate(native))])
29
+ expect(contents.last.dig('parts', 0, 'functionResponse')).to eq('name' => 'fixture_tool', 'response' => { 'ok' => true })
30
+ expect(Marshal.dump(history)).to eq(original)
31
+ end
32
+ end
33
+ end
@@ -0,0 +1,43 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+
5
+ RSpec.describe PWN::AI::Grok, 'reasoning effort wire' do
6
+ let(:completion) { { choices: [{ message: { role: 'assistant', content: 'ok' } }] }.to_json }
7
+
8
+ before do
9
+ allow(PWN::AI::HttpRetry).to receive(:provider_config).and_return(
10
+ model: 'grok-4.6', temp: 1, system_role_content: 'fixture', max_prompt_length: 256_000
11
+ )
12
+ allow(PWN::AI::Agent::PromptCache).to receive(:enabled?).and_return(false)
13
+ allow(described_class).to receive(:grok_rest_call).and_return(completion)
14
+ end
15
+
16
+ it 'maps reasoning_effort none to low for grok-4.6 chat' do
17
+ described_class.chat(request: 'title', model: 'grok-4.6', reasoning_effort: 'none', quiet: true, spinner: false)
18
+ expect(described_class).to have_received(:grok_rest_call).with(
19
+ hash_including(http_body: hash_including(model: 'grok-4.6', reasoning_effort: 'low'))
20
+ )
21
+ end
22
+
23
+ it 'maps reasoning_effort none to low for grok-4.7 chat' do
24
+ described_class.chat(request: 'title', model: 'grok-4.7', reasoning_effort: 'none', quiet: true, spinner: false)
25
+ expect(described_class).to have_received(:grok_rest_call).with(
26
+ hash_including(http_body: hash_including(model: 'grok-4.7', reasoning_effort: 'low'))
27
+ )
28
+ end
29
+
30
+ it 'omits reasoning_effort none for a Grok model with no effort menu' do
31
+ described_class.chat(request: 'title', model: 'grok-4', reasoning_effort: 'none', quiet: true, spinner: false)
32
+ expect(described_class).to have_received(:grok_rest_call).with(
33
+ hash_including(http_body: satisfy { |body| !body.key?(:reasoning_effort) })
34
+ )
35
+ end
36
+
37
+ it 'keeps an explicit supported effort' do
38
+ described_class.chat(request: 'title', model: 'grok-4.6', reasoning_effort: 'high', quiet: true, spinner: false)
39
+ expect(described_class).to have_received(:grok_rest_call).with(
40
+ hash_including(http_body: hash_including(reasoning_effort: 'high'))
41
+ )
42
+ end
43
+ end
@@ -0,0 +1,36 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+
5
+ RSpec.describe PWN::AI::MCP, 'backend-bound sessions' do
6
+ before do
7
+ described_class.reset!
8
+ %w[AuditOne AuditTwo].each do |name|
9
+ backend = Module.new
10
+ backend.define_singleton_method(:connect) { |_opts| {} }
11
+ backend.define_singleton_method(:disconnect) { |_opts| {} }
12
+ backend.define_singleton_method(:list_tools) { |_opts| [] }
13
+ backend.define_singleton_method(:call_tool) { |_opts| { parsed: 'fixture' } }
14
+ stub_const("PWN::AI::MCP::#{name}", backend)
15
+ end
16
+ described_class.connect(backend: 'audit_one', session_id: 'shared-fixture')
17
+ end
18
+
19
+ after { described_class.reset! }
20
+
21
+ %i[connect list_tools call_tool disconnect].each do |operation|
22
+ it "rejects #{operation} for a session belonging to another explicit backend" do
23
+ expect(PWN::AI::MCP::AuditOne).not_to receive(:call_tool)
24
+ expect(PWN::AI::MCP::AuditOne).not_to receive(:list_tools)
25
+ expect do
26
+ described_class.public_send(operation, backend: 'audit_two', session_id: 'shared-fixture', name: 'echo')
27
+ end.to raise_error(ArgumentError, /session.*audit_one.*audit_two/i)
28
+ expect(described_class.invoke(action: 'status', session_id: 'shared-fixture')).to include(connected: true, backend: 'audit_one')
29
+ end
30
+ end
31
+
32
+ it 'allows explicit force reconnect and leaves the new backend usable' do
33
+ expect(described_class.connect(backend: 'audit_two', session_id: 'shared-fixture', force: true)).to include(backend: 'audit_two')
34
+ expect(described_class.call_tool(session_id: 'shared-fixture', name: 'echo')).to include(backend: 'audit_two', parsed: 'fixture')
35
+ end
36
+ end
@@ -165,7 +165,7 @@ describe PWN::AI::Ollama do
165
165
  ]
166
166
  sse_lines = chunks.map { |c| "data: #{JSON.generate(c)}" }
167
167
  sse_lines << 'data: [DONE]'
168
- body = "#{sse_lines.join("\n")}\n"
168
+ body = "#{sse_lines.join("\n\n")}\n\n"
169
169
 
170
170
  out = described_class.send(
171
171
  :assemble_ollama_stream,
@@ -212,7 +212,7 @@ describe PWN::AI::Ollama do
212
212
  ]
213
213
  sse_lines = chunks.map { |c| "data: #{JSON.generate(c)}" }
214
214
  sse_lines << 'data: [DONE]'
215
- body = "#{sse_lines.join("\n")}\n"
215
+ body = "#{sse_lines.join("\n\n")}\n\n"
216
216
 
217
217
  out = described_class.send(
218
218
  :assemble_ollama_stream,
@@ -292,15 +292,10 @@ describe PWN::AI::Ollama do
292
292
  end.to raise_error(RuntimeError, /empty response|ERROR/)
293
293
  end
294
294
 
295
- it 'marks empty stream payloads instead of returning bare {}' do
296
- out = described_class.send(
297
- :assemble_ollama_stream,
298
- body: '',
299
- rest_call: 'ollama/api/chat'
300
- )
301
- parsed = JSON.parse(out, symbolize_names: true)
302
- expect(parsed[:done_reason]).to eq('error')
303
- expect(parsed.dig(:message, :error)).to eq('empty_stream_payloads')
295
+ it 'rejects empty stream payloads instead of inventing a terminal message' do
296
+ expect do
297
+ described_class.send(:assemble_ollama_stream, body: '', rest_call: 'ollama/api/chat')
298
+ end.to raise_error(RuntimeError, /incomplete stream.*no payloads/)
304
299
  end
305
300
 
306
301
  it 'normalizes string function.arguments into objects before ollama_rest_call' do
@@ -171,7 +171,7 @@ describe PWN::AI::OpenWebUI do
171
171
  ]
172
172
  sse_lines = chunks.map { |c| "data: #{JSON.generate(c)}" }
173
173
  sse_lines << 'data: [DONE]'
174
- body = "#{sse_lines.join("\n")}\n"
174
+ body = "#{sse_lines.join("\n\n")}\n\n"
175
175
 
176
176
  out = described_class.send(
177
177
  :assemble_openwebui_stream,
@@ -218,7 +218,7 @@ describe PWN::AI::OpenWebUI do
218
218
  ]
219
219
  sse_lines = chunks.map { |c| "data: #{JSON.generate(c)}" }
220
220
  sse_lines << 'data: [DONE]'
221
- body = "#{sse_lines.join("\n")}\n"
221
+ body = "#{sse_lines.join("\n\n")}\n\n"
222
222
 
223
223
  out = described_class.send(
224
224
  :assemble_openwebui_stream,
@@ -298,15 +298,10 @@ describe PWN::AI::OpenWebUI do
298
298
  end.to raise_error(RuntimeError, /empty response|ERROR/)
299
299
  end
300
300
 
301
- it 'marks empty stream payloads instead of returning bare {}' do
302
- out = described_class.send(
303
- :assemble_openwebui_stream,
304
- body: '',
305
- rest_call: 'api/chat'
306
- )
307
- parsed = JSON.parse(out, symbolize_names: true)
308
- expect(parsed[:done_reason]).to eq('error')
309
- expect(parsed.dig(:message, :error)).to eq('empty_stream_payloads')
301
+ it 'rejects empty stream payloads instead of inventing a terminal message' do
302
+ expect do
303
+ described_class.send(:assemble_openwebui_stream, body: '', rest_call: 'api/chat')
304
+ end.to raise_error(RuntimeError, /incomplete stream.*no payloads/)
310
305
  end
311
306
 
312
307
  it 'normalizes string function.arguments into objects before openwebui_rest_call' do
@@ -0,0 +1,44 @@
1
+ # frozen_string_literal: true
2
+
3
+ require 'spec_helper'
4
+ require 'rest-client'
5
+ require 'net/http'
6
+
7
+ { PWN::AI::OpenWebUI => :openwebui, PWN::AI::Ollama => :ollama }.each do |provider, key|
8
+ RSpec.describe provider, 'history without token usage' do
9
+ it 'continues the history returned by an SSE response with no usage frame' do
10
+ stub_const('PWN::Env', { ai: { key => { model: 'fixture-history', key: 'fixture-key', base_uri: 'https://example.test' } } })
11
+ allow(PWN::Plugins::TransparentBrowser).to receive(:open).and_return(browser: RestClient)
12
+ described_class.instance_variable_set(:@strict_system_cache, nil)
13
+ payloads = []
14
+ allow(RestClient::Request).to receive(:execute) do |args|
15
+ if args[:url].end_with?('/show')
16
+ '{}'
17
+ else
18
+ payloads << JSON.parse(args.fetch(:payload))
19
+ response = instance_double(Net::HTTPResponse, code: '200')
20
+ stream = "data: #{JSON.generate(choices: [{ delta: { role: 'assistant', content: 'Fixture answer' }, finish_reason: 'stop' }])}\n\ndata: [DONE]\n\n"
21
+ allow(response).to receive(:read_body).and_yield(stream)
22
+ args.fetch(:block_response).call(response)
23
+ end
24
+ end
25
+ first = described_class.chat(request: 'First fixture request')
26
+ expect(first).not_to have_key(:usage)
27
+ second = described_class.chat(request: 'Second fixture request', response_history: first)
28
+ expect(second[:choices].last).to include(role: 'assistant', content: 'Fixture answer')
29
+ expect(payloads.last.fetch('messages').map { |row| row['content'] }).to include('First fixture request', 'Fixture answer', 'Second fixture request')
30
+ ensure
31
+ described_class.instance_variable_set(:@strict_system_cache, nil)
32
+ end
33
+ end
34
+ end
35
+
36
+ RSpec.describe PWN::AI::Grok, 'quiet transport failure' do
37
+ it 'preserves transport exceptions instead of masking them with NoMethodError' do
38
+ stub_const('PWN::Env', { ai: { grok: { model: 'fixture', key: 'fixture-key' } } })
39
+ allow(PWN::Plugins::TransparentBrowser).to receive(:open).and_return(browser: RestClient)
40
+ failure = IOError.new('fixture transport failure')
41
+ allow(RestClient::Request).to receive(:execute).and_raise(failure)
42
+ expect { described_class.chat(request: 'fixture', quiet: true) }.to(raise_error { |error| expect(error).to equal(failure) })
43
+ end
44
+ end