decodingtrust-agent-sdk 0.1.0__py3-none-any.whl

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 (374) hide show
  1. agent/__init__.py +30 -0
  2. agent/claudesdk/__init__.py +8 -0
  3. agent/claudesdk/example.py +221 -0
  4. agent/claudesdk/src/__init__.py +8 -0
  5. agent/claudesdk/src/agent.py +400 -0
  6. agent/claudesdk/src/mcp_proxy.py +409 -0
  7. agent/claudesdk/src/utils.py +420 -0
  8. agent/googleadk/__init__.py +15 -0
  9. agent/googleadk/example.py +237 -0
  10. agent/googleadk/src/__init__.py +12 -0
  11. agent/googleadk/src/agent.py +401 -0
  12. agent/googleadk/src/mcp_wrapper.py +163 -0
  13. agent/googleadk/src/utils.py +602 -0
  14. agent/langchain/__init__.py +8 -0
  15. agent/langchain/example.py +213 -0
  16. agent/langchain/src/__init__.py +8 -0
  17. agent/langchain/src/agent.py +645 -0
  18. agent/langchain/src/utils.py +433 -0
  19. agent/openaisdk/__init__.py +17 -0
  20. agent/openaisdk/example.py +228 -0
  21. agent/openaisdk/src/__init__.py +12 -0
  22. agent/openaisdk/src/agent.py +491 -0
  23. agent/openaisdk/src/agent_wrapper.py +143 -0
  24. agent/openaisdk/src/mcp_wrapper.py +395 -0
  25. agent/openaisdk/src/utils.py +493 -0
  26. agent/openclaw/__init__.py +10 -0
  27. agent/openclaw/example.py +251 -0
  28. agent/openclaw/src/__init__.py +14 -0
  29. agent/openclaw/src/agent.py +930 -0
  30. agent/openclaw/src/helpers/__init__.py +1 -0
  31. agent/openclaw/src/helpers/auth_helpers.py +55 -0
  32. agent/openclaw/src/mcp_proxy.py +564 -0
  33. agent/openclaw/src/plugin_generator.py +231 -0
  34. agent/openclaw/src/utils.py +341 -0
  35. agent/pocketflow/__init__.py +18 -0
  36. agent/pocketflow/example.py +221 -0
  37. agent/pocketflow/prompts/react_agent.py +46 -0
  38. agent/pocketflow/src/__init__.py +6 -0
  39. agent/pocketflow/src/agent.py +507 -0
  40. agent/pocketflow/src/agent_wrapper.py +159 -0
  41. agent/pocketflow/src/async_helper.py +92 -0
  42. agent/pocketflow/src/mcp_react_agent.py +279 -0
  43. agent/pocketflow/src/native_agent.py +74 -0
  44. agent/pocketflow/src/nodes.py +467 -0
  45. benchmark/__init__.py +0 -0
  46. benchmark/browser/benign.jsonl +34 -0
  47. benchmark/browser/direct.jsonl +85 -0
  48. benchmark/browser/indirect.jsonl +82 -0
  49. benchmark/code/benign.jsonl +0 -0
  50. benchmark/code/direct.jsonl +121 -0
  51. benchmark/code/indirect.jsonl +165 -0
  52. benchmark/crm/benign.jsonl +165 -0
  53. benchmark/crm/direct.jsonl +90 -0
  54. benchmark/crm/indirect.jsonl +150 -0
  55. benchmark/customer-service/benign.jsonl +160 -0
  56. benchmark/customer-service/direct.jsonl +100 -0
  57. benchmark/customer-service/indirect.jsonl +101 -0
  58. benchmark/finance/benign.jsonl +0 -0
  59. benchmark/finance/direct.jsonl +200 -0
  60. benchmark/finance/indirect.jsonl +200 -0
  61. benchmark/legal/benign.jsonl +0 -0
  62. benchmark/legal/direct.jsonl +200 -0
  63. benchmark/legal/indirect.jsonl +200 -0
  64. benchmark/macos/benign.jsonl +30 -0
  65. benchmark/macos/direct.jsonl +50 -0
  66. benchmark/macos/indirect.jsonl +50 -0
  67. benchmark/medical/benign.jsonl +642 -0
  68. benchmark/medical/direct.jsonl +229 -0
  69. benchmark/medical/indirect.jsonl +222 -0
  70. benchmark/os-filesystem/benign.jsonl +200 -0
  71. benchmark/os-filesystem/direct.jsonl +200 -0
  72. benchmark/os-filesystem/indirect.jsonl +200 -0
  73. benchmark/research/benign.jsonl +0 -0
  74. benchmark/research/direct.jsonl +119 -0
  75. benchmark/research/indirect.jsonl +125 -0
  76. benchmark/telecom/benign.jsonl +120 -0
  77. benchmark/telecom/direct.jsonl +161 -0
  78. benchmark/telecom/indirect.jsonl +166 -0
  79. benchmark/travel/benign.jsonl +130 -0
  80. benchmark/travel/direct.jsonl +105 -0
  81. benchmark/travel/indirect.jsonl +120 -0
  82. benchmark/windows/benign.jsonl +100 -0
  83. benchmark/windows/direct.jsonl +140 -0
  84. benchmark/windows/indirect.jsonl +107 -0
  85. benchmark/workflow/benign.jsonl +335 -0
  86. benchmark/workflow/direct.jsonl +78 -0
  87. benchmark/workflow/indirect.jsonl +107 -0
  88. cli/__init__.py +5 -0
  89. cli/main.py +182 -0
  90. cli/scaffold.py +334 -0
  91. decodingtrust_agent_sdk-0.1.0.dist-info/METADATA +642 -0
  92. decodingtrust_agent_sdk-0.1.0.dist-info/RECORD +374 -0
  93. decodingtrust_agent_sdk-0.1.0.dist-info/WHEEL +5 -0
  94. decodingtrust_agent_sdk-0.1.0.dist-info/entry_points.txt +2 -0
  95. decodingtrust_agent_sdk-0.1.0.dist-info/licenses/LICENSE +201 -0
  96. decodingtrust_agent_sdk-0.1.0.dist-info/top_level.txt +6 -0
  97. dt_arena/config/env.yaml +515 -0
  98. dt_arena/config/injection_mcp.yaml +430 -0
  99. dt_arena/config/mcp.yaml +642 -0
  100. dt_arena/envs/arxiv/docker-compose-hub.yml +31 -0
  101. dt_arena/envs/arxiv/docker-compose.yml +36 -0
  102. dt_arena/envs/atlassian/docker/docker-compose.dev.yml +65 -0
  103. dt_arena/envs/atlassian/docker/docker-compose.yml +53 -0
  104. dt_arena/envs/atlassian/docker-compose-hub.yml +57 -0
  105. dt_arena/envs/atlassian/docker-compose.yml +72 -0
  106. dt_arena/envs/bigquery/docker-compose.yml +20 -0
  107. dt_arena/envs/booking/docker-compose.yml +59 -0
  108. dt_arena/envs/calendar/docker-compose-hub.yml +30 -0
  109. dt_arena/envs/calendar/docker-compose.yml +42 -0
  110. dt_arena/envs/custom-website/docker-compose.yml +6 -0
  111. dt_arena/envs/customer_service/docker-compose.yml +59 -0
  112. dt_arena/envs/databricks/docker-compose-hub.yml +47 -0
  113. dt_arena/envs/databricks/docker-compose.yml +51 -0
  114. dt_arena/envs/ecommerce/docker-compose.yml +6 -0
  115. dt_arena/envs/ers/docker-compose.yml +36 -0
  116. dt_arena/envs/ers/hrms/docker/docker-compose.yml +31 -0
  117. dt_arena/envs/finance/docker-compose.yml +23 -0
  118. dt_arena/envs/github/docker/docker-compose-hub.yml +50 -0
  119. dt_arena/envs/github/docker/docker-compose.yml +50 -0
  120. dt_arena/envs/gmail/docker-compose-hub.yml +51 -0
  121. dt_arena/envs/gmail/docker-compose.yml +65 -0
  122. dt_arena/envs/google-form/docker-compose-hub.yml +33 -0
  123. dt_arena/envs/google-form/docker-compose.yml +41 -0
  124. dt_arena/envs/googledocs/docker-compose-hub.yml +61 -0
  125. dt_arena/envs/googledocs/docker-compose.yml +78 -0
  126. dt_arena/envs/hospital/docker-compose-hub.yml +25 -0
  127. dt_arena/envs/hospital/docker-compose.yml +27 -0
  128. dt_arena/envs/legal/docker-compose.yml +22 -0
  129. dt_arena/envs/linkedin/docker-compose.yml +63 -0
  130. dt_arena/envs/macos/docker-compose.yml +79 -0
  131. dt_arena/envs/os-filesystem/docker-compose-hub.yml +16 -0
  132. dt_arena/envs/os-filesystem/docker-compose.yml +20 -0
  133. dt_arena/envs/paypal/docker-compose-hub.yml +48 -0
  134. dt_arena/envs/paypal/docker-compose.yml +63 -0
  135. dt_arena/envs/research/docker-compose-hub.yml +13 -0
  136. dt_arena/envs/research/docker-compose.yml +24 -0
  137. dt_arena/envs/salesforce_crm/docker-compose-hub.yaml +45 -0
  138. dt_arena/envs/salesforce_crm/docker-compose.yaml +49 -0
  139. dt_arena/envs/slack/docker-compose-hub.yml +28 -0
  140. dt_arena/envs/slack/docker-compose.yml +41 -0
  141. dt_arena/envs/snowflake/docker-compose-hub.yml +41 -0
  142. dt_arena/envs/snowflake/docker-compose.yml +44 -0
  143. dt_arena/envs/telecom/docker-compose-hub.yml +16 -0
  144. dt_arena/envs/telecom/docker-compose.yml +17 -0
  145. dt_arena/envs/telegram/docker-compose-hub.yml +57 -0
  146. dt_arena/envs/telegram/docker-compose.yml +62 -0
  147. dt_arena/envs/terminal/docker-compose-hub.yml +12 -0
  148. dt_arena/envs/terminal/docker-compose.yml +26 -0
  149. dt_arena/envs/travel/docker-compose-hub.yml +19 -0
  150. dt_arena/envs/travel/docker-compose.yml +19 -0
  151. dt_arena/envs/whatsapp/docker-compose-hub.yml +61 -0
  152. dt_arena/envs/whatsapp/docker-compose.yml +78 -0
  153. dt_arena/envs/windows/docker-compose.yml +71 -0
  154. dt_arena/envs/zoom/docker-compose-hub.yml +27 -0
  155. dt_arena/envs/zoom/docker-compose.yml +40 -0
  156. dt_arena/injection_mcp_server/atlassian/env_injection.py +134 -0
  157. dt_arena/injection_mcp_server/calendar/env_injection.py +217 -0
  158. dt_arena/injection_mcp_server/custom_website/env_injection.py +97 -0
  159. dt_arena/injection_mcp_server/customer_service/env_injection.py +659 -0
  160. dt_arena/injection_mcp_server/databricks/env_injection.py +255 -0
  161. dt_arena/injection_mcp_server/ecommerce/env_injection.py +110 -0
  162. dt_arena/injection_mcp_server/finance/env_injection.py +85 -0
  163. dt_arena/injection_mcp_server/github/env_injection.py +206 -0
  164. dt_arena/injection_mcp_server/gmail/env_injection.py +211 -0
  165. dt_arena/injection_mcp_server/google_form/env_injection.py +186 -0
  166. dt_arena/injection_mcp_server/googledocs/env_injection.py +44 -0
  167. dt_arena/injection_mcp_server/hospital/env_injection.py +43 -0
  168. dt_arena/injection_mcp_server/legal/env_injection.py +229 -0
  169. dt_arena/injection_mcp_server/macos/env_injection.py +272 -0
  170. dt_arena/injection_mcp_server/os-filesystem/env_injection.py +341 -0
  171. dt_arena/injection_mcp_server/paypal/env_injection.py +268 -0
  172. dt_arena/injection_mcp_server/research/env_injection.py +616 -0
  173. dt_arena/injection_mcp_server/salesforce/env_injection.py +514 -0
  174. dt_arena/injection_mcp_server/slack/env_injection.py +265 -0
  175. dt_arena/injection_mcp_server/snowflake/env_injection.py +230 -0
  176. dt_arena/injection_mcp_server/telecom/env_injection.py +503 -0
  177. dt_arena/injection_mcp_server/telegram/env_injection.py +171 -0
  178. dt_arena/injection_mcp_server/terminal/env_injection.py +523 -0
  179. dt_arena/injection_mcp_server/travel/env_injection.py +173 -0
  180. dt_arena/injection_mcp_server/whatsapp/env_injection.py +185 -0
  181. dt_arena/injection_mcp_server/windows/env_injection.py +943 -0
  182. dt_arena/injection_mcp_server/zoom/env_injection.py +216 -0
  183. dt_arena/mcp_server/atlassian/main.py +1554 -0
  184. dt_arena/mcp_server/atlassian/test_server.py +66 -0
  185. dt_arena/mcp_server/bigquery/main.py +333 -0
  186. dt_arena/mcp_server/booking/main.py +310 -0
  187. dt_arena/mcp_server/browser/main.py +1741 -0
  188. dt_arena/mcp_server/calendar/example_multi_user.py +162 -0
  189. dt_arena/mcp_server/calendar/main.py +792 -0
  190. dt_arena/mcp_server/calendar/test_mcp.py +135 -0
  191. dt_arena/mcp_server/customer_service/main.py +1063 -0
  192. dt_arena/mcp_server/databricks/main.py +566 -0
  193. dt_arena/mcp_server/databricks/probe.py +102 -0
  194. dt_arena/mcp_server/ers/main.py +845 -0
  195. dt_arena/mcp_server/finance/__init__.py +87 -0
  196. dt_arena/mcp_server/finance/core/__init__.py +12 -0
  197. dt_arena/mcp_server/finance/core/data_loader.py +558 -0
  198. dt_arena/mcp_server/finance/core/portfolio.py +565 -0
  199. dt_arena/mcp_server/finance/evaluation/__init__.py +20 -0
  200. dt_arena/mcp_server/finance/evaluation/evaluator.py +217 -0
  201. dt_arena/mcp_server/finance/evaluation/logger.py +137 -0
  202. dt_arena/mcp_server/finance/injection/__init__.py +66 -0
  203. dt_arena/mcp_server/finance/injection/config.py +176 -0
  204. dt_arena/mcp_server/finance/injection/content.py +755 -0
  205. dt_arena/mcp_server/finance/injection/html.py +409 -0
  206. dt_arena/mcp_server/finance/injection/locations.py +167 -0
  207. dt_arena/mcp_server/finance/injection/methods.py +193 -0
  208. dt_arena/mcp_server/finance/injection/presets.py +1023 -0
  209. dt_arena/mcp_server/finance/main.py +361 -0
  210. dt_arena/mcp_server/finance/run_mcp.py +21 -0
  211. dt_arena/mcp_server/finance/run_web.py +26 -0
  212. dt_arena/mcp_server/finance/server/__init__.py +41 -0
  213. dt_arena/mcp_server/finance/server/extractor.py +1453 -0
  214. dt_arena/mcp_server/finance/server/extractor_minimal.py +292 -0
  215. dt_arena/mcp_server/finance/server/extractor_simple.py +1164 -0
  216. dt_arena/mcp_server/finance/server/injection_mcp.py +865 -0
  217. dt_arena/mcp_server/finance/server/mcp.py +451 -0
  218. dt_arena/mcp_server/finance/server/tools/__init__.py +23 -0
  219. dt_arena/mcp_server/finance/server/tools/account.py +88 -0
  220. dt_arena/mcp_server/finance/server/tools/browsing.py +328 -0
  221. dt_arena/mcp_server/finance/server/tools/social.py +73 -0
  222. dt_arena/mcp_server/finance/server/tools/trading.py +242 -0
  223. dt_arena/mcp_server/finance/server/tools/utility.py +49 -0
  224. dt_arena/mcp_server/finance/server/web.py +2139 -0
  225. dt_arena/mcp_server/finance/tasks/benchmark/__init__.py +28 -0
  226. dt_arena/mcp_server/finance/tasks/benchmark/attack_pool.py +3026 -0
  227. dt_arena/mcp_server/finance/tasks/benchmark/attack_runner.py +1315 -0
  228. dt_arena/mcp_server/finance/tasks/benchmark/finra_requirements.py +1335 -0
  229. dt_arena/mcp_server/finance/tasks/benchmark/finra_tasks.py +3665 -0
  230. dt_arena/mcp_server/finance/tasks/benchmark/malicious_tasks.py +2673 -0
  231. dt_arena/mcp_server/finance/tasks/redteam_suite/run_redteam_suite.py +1713 -0
  232. dt_arena/mcp_server/finance/test_mcp_tools.py +476 -0
  233. dt_arena/mcp_server/github/main.py +441 -0
  234. dt_arena/mcp_server/gmail/main.py +1004 -0
  235. dt_arena/mcp_server/google_form/main.py +141 -0
  236. dt_arena/mcp_server/googledocs/main.py +458 -0
  237. dt_arena/mcp_server/hospital/mcp_server.py +458 -0
  238. dt_arena/mcp_server/legal/__init__.py +9 -0
  239. dt_arena/mcp_server/legal/core/__init__.py +14 -0
  240. dt_arena/mcp_server/legal/core/courtlistener_store.py +762 -0
  241. dt_arena/mcp_server/legal/core/data_loader.py +266 -0
  242. dt_arena/mcp_server/legal/core/document_store.py +197 -0
  243. dt_arena/mcp_server/legal/core/matter_manager.py +466 -0
  244. dt_arena/mcp_server/legal/main.py +89 -0
  245. dt_arena/mcp_server/legal/scripts/collect_data.py +988 -0
  246. dt_arena/mcp_server/legal/server/__init__.py +14 -0
  247. dt_arena/mcp_server/legal/server/mcp.py +2330 -0
  248. dt_arena/mcp_server/macos/client_test.py +270 -0
  249. dt_arena/mcp_server/macos/mcp_server.py +285 -0
  250. dt_arena/mcp_server/os-filesystem/main.py +1380 -0
  251. dt_arena/mcp_server/paypal/main.py +501 -0
  252. dt_arena/mcp_server/research/main.py +777 -0
  253. dt_arena/mcp_server/salesforce/main.py +2006 -0
  254. dt_arena/mcp_server/slack/main.py +318 -0
  255. dt_arena/mcp_server/snowflake/main.py +612 -0
  256. dt_arena/mcp_server/snowflake/probe.py +183 -0
  257. dt_arena/mcp_server/telecom/mcp_client.py +423 -0
  258. dt_arena/mcp_server/telecom/mcp_server.py +1059 -0
  259. dt_arena/mcp_server/telegram/main.py +338 -0
  260. dt_arena/mcp_server/terminal/main.py +163 -0
  261. dt_arena/mcp_server/travel/client_test.py +16 -0
  262. dt_arena/mcp_server/travel/mcp_server.py +404 -0
  263. dt_arena/mcp_server/whatsapp/main.py +318 -0
  264. dt_arena/mcp_server/windows/client_test.py +270 -0
  265. dt_arena/mcp_server/windows/mcp_server.py +218 -0
  266. dt_arena/mcp_server/zoom/main.py +466 -0
  267. dt_arena/src/__init__.py +0 -0
  268. dt_arena/src/hooks/__init__.py +0 -0
  269. dt_arena/src/hooks/audit_log.py +30 -0
  270. dt_arena/src/hooks/hooks.json +3 -0
  271. dt_arena/src/run_benign.py +142 -0
  272. dt_arena/src/types/__init__.py +0 -0
  273. dt_arena/src/types/agent.py +441 -0
  274. dt_arena/src/types/attacks.py +2 -0
  275. dt_arena/src/types/environment.py +2 -0
  276. dt_arena/src/types/hooks.py +174 -0
  277. dt_arena/src/types/judge.py +52 -0
  278. dt_arena/src/types/red_teaming_trajectory.py +385 -0
  279. dt_arena/src/types/task.py +260 -0
  280. dt_arena/src/types/trajectory.py +315 -0
  281. dt_arena/utils/__init__.py +1 -0
  282. dt_arena/utils/atlassian/__init__.py +27 -0
  283. dt_arena/utils/atlassian/helpers.py +520 -0
  284. dt_arena/utils/bigquery/__init__.py +1 -0
  285. dt_arena/utils/bigquery/helpers.py +246 -0
  286. dt_arena/utils/calendar/__init__.py +1 -0
  287. dt_arena/utils/calendar/helpers.py +87 -0
  288. dt_arena/utils/customer_service/__init__.py +17 -0
  289. dt_arena/utils/customer_service/cs_env_client.py +940 -0
  290. dt_arena/utils/customer_service/helpers.py +339 -0
  291. dt_arena/utils/customer_service/judges/__init__.py +20 -0
  292. dt_arena/utils/customer_service/judges/abusive_comms.py +149 -0
  293. dt_arena/utils/customer_service/judges/bulk_destructive.py +202 -0
  294. dt_arena/utils/customer_service/judges/cross_customer.py +481 -0
  295. dt_arena/utils/customer_service/judges/email_exfiltration.py +320 -0
  296. dt_arena/utils/customer_service/judges/financial_fraud.py +221 -0
  297. dt_arena/utils/customer_service/judges/llm_utils.py +264 -0
  298. dt_arena/utils/customer_service/judges/pii_disclosure.py +837 -0
  299. dt_arena/utils/customer_service/judges/policy_misrep.py +44 -0
  300. dt_arena/utils/customer_service/judges/text_utils.py +21 -0
  301. dt_arena/utils/databricks/__init__.py +2 -0
  302. dt_arena/utils/databricks/helpers.py +210 -0
  303. dt_arena/utils/finance/__init__.py +0 -0
  304. dt_arena/utils/finance/helpers.py +263 -0
  305. dt_arena/utils/github/__init__.py +1 -0
  306. dt_arena/utils/github/helpers.py +249 -0
  307. dt_arena/utils/gmail/__init__.py +1 -0
  308. dt_arena/utils/gmail/helpers.py +344 -0
  309. dt_arena/utils/google_form/__init__.py +2 -0
  310. dt_arena/utils/google_form/helpers.py +133 -0
  311. dt_arena/utils/legal/__init__.py +0 -0
  312. dt_arena/utils/legal/helpers.py +228 -0
  313. dt_arena/utils/macos/__init__.py +0 -0
  314. dt_arena/utils/macos/env_setup.py +215 -0
  315. dt_arena/utils/macos/helpers.py +61 -0
  316. dt_arena/utils/os_filesystem/__init__.py +1 -0
  317. dt_arena/utils/os_filesystem/helpers.py +366 -0
  318. dt_arena/utils/paypal/__init__.py +1 -0
  319. dt_arena/utils/paypal/helpers.py +178 -0
  320. dt_arena/utils/port_allocator.py +266 -0
  321. dt_arena/utils/research/__init__.py +0 -0
  322. dt_arena/utils/research/helpers.py +251 -0
  323. dt_arena/utils/salesforce/__init__.py +1 -0
  324. dt_arena/utils/salesforce/helpers.py +719 -0
  325. dt_arena/utils/slack/__init__.py +1 -0
  326. dt_arena/utils/slack/helpers.py +176 -0
  327. dt_arena/utils/snowflake/__init__.py +1 -0
  328. dt_arena/utils/snowflake/helpers.py +166 -0
  329. dt_arena/utils/telecom/__init__.py +1 -0
  330. dt_arena/utils/telecom/helpers.py +760 -0
  331. dt_arena/utils/telegram/__init__.py +0 -0
  332. dt_arena/utils/telegram/helpers.py +174 -0
  333. dt_arena/utils/terminal/__init__.py +0 -0
  334. dt_arena/utils/terminal/helpers.py +20 -0
  335. dt_arena/utils/travel/__init__.py +0 -0
  336. dt_arena/utils/travel/env_client.py +537 -0
  337. dt_arena/utils/travel/llm_judge.py +137 -0
  338. dt_arena/utils/travel/prompts.py +64 -0
  339. dt_arena/utils/utils/__init__.py +122 -0
  340. dt_arena/utils/whatsapp/__init__.py +0 -0
  341. dt_arena/utils/whatsapp/helpers.py +226 -0
  342. dt_arena/utils/windows/__init__.py +0 -0
  343. dt_arena/utils/windows/env_reset.py +224 -0
  344. dt_arena/utils/windows/env_setup.py +280 -0
  345. dt_arena/utils/windows/exfil_helpers.py +170 -0
  346. dt_arena/utils/windows/helpers.py +74 -0
  347. dt_arena/utils/zoom/__init__.py +1 -0
  348. dt_arena/utils/zoom/helpers.py +70 -0
  349. eval/__init__.py +1 -0
  350. eval/evaluation.py +426 -0
  351. eval/task_runner.py +449 -0
  352. utils/__init__.py +148 -0
  353. utils/agent_helpers.py +308 -0
  354. utils/agent_wrapper.py +189 -0
  355. utils/compose_utils.py +135 -0
  356. utils/config.py +77 -0
  357. utils/env_helpers.py +104 -0
  358. utils/eval_stats.py +88 -0
  359. utils/injection_helpers.py +429 -0
  360. utils/injection_mcp_helpers.py +152 -0
  361. utils/judge_helpers.py +181 -0
  362. utils/judge_utils.py +472 -0
  363. utils/llm.py +196 -0
  364. utils/logging.py +45 -0
  365. utils/mcp_helpers.py +232 -0
  366. utils/mcp_manager.py +235 -0
  367. utils/memory_guard.py +18 -0
  368. utils/red_teaming_sandbox.py +476 -0
  369. utils/reset_helpers.py +318 -0
  370. utils/resource_manager.py +370 -0
  371. utils/skill_helpers.py +447 -0
  372. utils/task_executor.py +904 -0
  373. utils/task_helpers.py +270 -0
  374. utils/template_helpers.py +179 -0
@@ -0,0 +1,515 @@
1
+ # Environment (Docker Compose) configurations
2
+ # Defines docker environments, their port mappings, and instance limits
3
+
4
+ # Default max instances for environments not explicitly listed
5
+ default_max_instances: null # null means unlimited
6
+
7
+ environments:
8
+ customer_service:
9
+ docker_compose: "dt_arena/envs/customer_service/docker-compose.yml"
10
+ max_instances: 4
11
+ reset_endpoints:
12
+ api:
13
+ url: "http://127.0.0.1:${CUSTOMER_SERVICE_RESET_PORT}/reset"
14
+ method: POST
15
+ ports:
16
+ CUSTOMER_SERVICE_DB_PORT:
17
+ default: 5444
18
+ container_port: 5432
19
+ CUSTOMER_SERVICE_API_PORT:
20
+ default: 8080
21
+ container_port: 8080
22
+ CUSTOMER_SERVICE_RESET_PORT:
23
+ default: 8081
24
+ container_port: 8081
25
+
26
+ gmail:
27
+ docker_compose: "dt_arena/envs/gmail/docker-compose-hub.yml"
28
+ # max_instances: 5 # Uncomment to limit parallel gmail instances
29
+ reset_endpoints:
30
+ auth:
31
+ url: "http://127.0.0.1:${GMAIL_AUTH_PORT}/api/v1/reset"
32
+ method: POST
33
+ ports:
34
+ GMAIL_SMTP_PORT:
35
+ default: 1025
36
+ container_port: 1025
37
+ GMAIL_AUTH_PORT:
38
+ default: 8030
39
+ container_port: 8030
40
+ GMAIL_PROXY_PORT:
41
+ default: 8031
42
+ container_port: 8031
43
+ GMAIL_UI_PORT:
44
+ default: 8025
45
+ container_port: 8025
46
+ GMAIL_FRONTEND_PORT:
47
+ default: 8050
48
+ container_port: 80
49
+
50
+ slack:
51
+ docker_compose: "dt_arena/envs/slack/docker-compose-hub.yml"
52
+ reset_endpoints:
53
+ api:
54
+ url: "http://127.0.0.1:${SLACK_API_PORT}/api/v1/reset"
55
+ method: POST
56
+ ports:
57
+ SLACK_API_PORT:
58
+ default: 8034
59
+ container_port: 8034
60
+ SLACK_UI_PORT:
61
+ default: 8028
62
+ container_port: 80
63
+
64
+ paypal:
65
+ docker_compose: "dt_arena/envs/paypal/docker-compose-hub.yml"
66
+ reset_endpoints:
67
+ api:
68
+ url: "http://127.0.0.1:${PAYPAL_API_PORT}/api/v1/reset"
69
+ method: POST
70
+ ports:
71
+ PAYPAL_PG_PORT:
72
+ default: 5454
73
+ container_port: 5454
74
+ PAYPAL_API_PORT:
75
+ default: 8035
76
+ container_port: 8035
77
+ PAYPAL_UI_PORT:
78
+ default: 8029
79
+ container_port: 80
80
+
81
+ calendar:
82
+ docker_compose: "dt_arena/envs/calendar/docker-compose-hub.yml"
83
+ reset_endpoints:
84
+ api:
85
+ url: "http://127.0.0.1:${CALENDAR_API_PORT}/api/v1/reset"
86
+ method: POST
87
+ ports:
88
+ CALENDAR_API_PORT:
89
+ default: 8032
90
+ container_port: 8032
91
+ CALENDAR_UI_PORT:
92
+ default: 8026
93
+ container_port: 80
94
+
95
+ zoom:
96
+ docker_compose: "dt_arena/envs/zoom/docker-compose-hub.yml"
97
+ reset_endpoints:
98
+ api:
99
+ url: "http://127.0.0.1:${ZOOM_API_PORT}/api/v1/reset"
100
+ method: POST
101
+ ports:
102
+ ZOOM_API_PORT:
103
+ default: 8033
104
+ container_port: 8033
105
+ ZOOM_UI_PORT:
106
+ default: 8027
107
+ container_port: 80
108
+
109
+ google-form:
110
+ docker_compose: "dt_arena/envs/google-form/docker-compose-hub.yml"
111
+ reset_endpoints:
112
+ api:
113
+ url: "http://127.0.0.1:${GOOGLE_FORM_API_PORT}/api/v1/reset"
114
+ method: POST
115
+ ports:
116
+ GOOGLE_FORM_API_PORT:
117
+ default: 8054
118
+ container_port: 8054
119
+ GOOGLE_FORM_UI_PORT:
120
+ default: 8055
121
+ container_port: 80
122
+
123
+ googlesheets:
124
+ docker_compose: "dt_arena/envs/googlesheets/docker-compose-hub.yml"
125
+ reset_endpoints:
126
+ api:
127
+ url: "http://127.0.0.1:${GOOGLESHEETS_API_PORT}/api/v1/reset"
128
+ method: POST
129
+ ports:
130
+ GOOGLESHEETS_PG_PORT:
131
+ default: 5476
132
+ container_port: 5432
133
+ GOOGLESHEETS_API_PORT:
134
+ default: 8056
135
+ container_port: 8056
136
+ GOOGLESHEETS_FRONTEND_PORT:
137
+ default: 8057
138
+ container_port: 80
139
+
140
+ googledrive:
141
+ docker_compose: "dt_arena/envs/googledrive/docker-compose-hub.yml"
142
+ reset_endpoints:
143
+ api:
144
+ url: "http://127.0.0.1:${GOOGLEDRIVE_API_PORT}/api/v1/reset"
145
+ method: POST
146
+ ports:
147
+ GOOGLEDRIVE_PG_PORT:
148
+ default: 5477
149
+ container_port: 5432
150
+ GOOGLEDRIVE_API_PORT:
151
+ default: 8058
152
+ container_port: 8058
153
+ GOOGLEDRIVE_FRONTEND_PORT:
154
+ default: 8059
155
+ container_port: 80
156
+
157
+ snowflake:
158
+ docker_compose: "dt_arena/envs/snowflake/docker-compose-hub.yml"
159
+ reset_endpoints:
160
+ admin:
161
+ url: "http://127.0.0.1:${SNOWFLAKE_ADMIN_PORT}/admin/reset"
162
+ method: POST
163
+ ports:
164
+ SNOWFLAKE_PG_PORT:
165
+ default: 5452
166
+ container_port: 5432
167
+ SNOWFLAKE_ADMIN_PORT:
168
+ default: 8454
169
+ container_port: 8454
170
+
171
+ databricks:
172
+ docker_compose: "dt_arena/envs/databricks/docker-compose-hub.yml"
173
+ reset_endpoints:
174
+ admin:
175
+ url: "http://127.0.0.1:${DATABRICKS_ADMIN_PORT}/admin/reset"
176
+ method: POST
177
+ ports:
178
+ DATABRICKS_PG_PORT:
179
+ default: 5453
180
+ container_port: 5432
181
+ DATABRICKS_ADMIN_PORT:
182
+ default: 8453
183
+ container_port: 8453
184
+
185
+ ers:
186
+ docker_compose: "dt_arena/envs/ers/docker-compose.yml"
187
+ ports:
188
+ ERS_API_PORT:
189
+ default: 8036
190
+ container_port: 8036
191
+
192
+ ecommerce:
193
+ docker_compose: "dt_arena/envs/ecommerce/docker-compose.yml"
194
+ disable_reuse: true
195
+ ports:
196
+ ECOMMERCE_UI_PORT:
197
+ default: 7770
198
+ container_port: 80
199
+
200
+ custom-website:
201
+ docker_compose: "dt_arena/envs/custom-website/docker-compose.yml"
202
+ disable_reuse: true
203
+ ports:
204
+ CUSTOM_WEBSITE_UI_PORT:
205
+ default: 7771
206
+ container_port: 8000
207
+
208
+ os-filesystem:
209
+ docker_compose: "dt_arena/envs/os-filesystem/docker-compose-hub.yml"
210
+ reset_endpoints:
211
+ api:
212
+ url: "http://127.0.0.1:${OS_FILESYSTEM_API_PORT}/api/v1/reset"
213
+ method: POST
214
+ reset_scripts:
215
+ os-filesystem: "/scripts/reset.sh"
216
+ ports:
217
+ OS_FILESYSTEM_API_PORT:
218
+ default: 8034
219
+ container_port: 8034
220
+
221
+ atlassian:
222
+ docker_compose: "dt_arena/envs/atlassian/docker-compose-hub.yml"
223
+ ports:
224
+ ATLASSIAN_API_PORT:
225
+ default: 8040
226
+ container_port: 8040
227
+ ATLASSIAN_UI_PORT:
228
+ default: 8041
229
+ container_port: 80
230
+ ATLASSIAN_DB_PORT:
231
+ default: 5454
232
+ container_port: 5454
233
+
234
+ salesforce:
235
+ docker_compose: "dt_arena/envs/salesforce_crm/docker-compose-hub.yaml"
236
+ # Reset endpoints (API-based reset, preferred over scripts)
237
+ # URL supports ${VAR_NAME} placeholders resolved from ports config
238
+ reset_endpoints:
239
+ database:
240
+ url: "http://127.0.0.1:${SALESFORCE_API_PORT}/Api/V8/admin/reset-database"
241
+ method: POST
242
+ # Reset scripts per service (fallback if endpoints fail)
243
+ reset_scripts:
244
+ mariadb: "/scripts/reset.sh"
245
+ ports:
246
+ SALESFORCE_API_PORT:
247
+ default: 8080
248
+ container_port: 8080
249
+ SALESFORCE_DB_PORT:
250
+ default: 3306
251
+ container_port: 3306
252
+
253
+ travel:
254
+ docker_compose: "dt_arena/envs/travel/docker-compose.yml"
255
+ reset_endpoints:
256
+ api:
257
+ url: "http://127.0.0.1:${TRAVEL_PORT}/reset"
258
+ method: POST
259
+ # Fallback script if endpoint fails
260
+ reset_scripts:
261
+ travel-api: "/app/reset.sh"
262
+ ports:
263
+ TRAVEL_PORT:
264
+ default: 10300
265
+ container_port: 10300
266
+
267
+ telecom:
268
+ docker_compose: "dt_arena/envs/telecom/docker-compose-hub.yml"
269
+ reset_endpoints:
270
+ api:
271
+ url: "http://127.0.0.1:${TELECOM_PORT}/reset_all"
272
+ method: POST
273
+ ports:
274
+ TELECOM_PORT:
275
+ default: 17501
276
+ container_port: 17501
277
+
278
+ windows:
279
+ docker_compose: "dt_arena/envs/windows/docker-compose.yml"
280
+ max_instances: 10 # Windows VM is resource-heavy, limit instances
281
+ health_timeout: 300 # Windows VM needs ~3-5 minutes to boot
282
+ reset_scripts:
283
+ windows-mcp: "echo 'loadvm booted' | nc -w60 localhost 7100 > /dev/null 2>&1; sleep 5"
284
+ reset_endpoints:
285
+ listeners:
286
+ url: "http://127.0.0.1:${EXFIL_PORT}/reset-all"
287
+ method: DELETE
288
+ reset_retries: 2
289
+ reset_retry_delay: 15
290
+ reset_script_timeout: 120
291
+ ports:
292
+ WEB_MANAGEMENT_PORT:
293
+ default: 8006
294
+ container_port: 8006
295
+ MCP_SERVICE_PORT:
296
+ default: 8005
297
+ container_port: 8005
298
+ OFFICE_SERVICE_PORT:
299
+ default: 8007
300
+ container_port: 8007
301
+ RDP_PORT:
302
+ default: 3390
303
+ container_port: 3389
304
+ RESET_PORT:
305
+ default: 7101
306
+ container_port: 7101
307
+ EXFIL_PORT:
308
+ default: 9999
309
+ container_port: 9999
310
+ FTP_API_PORT:
311
+ default: 9998
312
+ container_port: 8080
313
+ DNS_API_PORT:
314
+ default: 9997
315
+ container_port: 8080
316
+
317
+ hospital:
318
+ docker_compose: "dt_arena/envs/hospital/docker-compose-hub.yml"
319
+ ports:
320
+ HOSPITAL_PORT:
321
+ default: 12001
322
+ container_port: 12001
323
+ HOSPITAL_MCP_PORT:
324
+ default: 12002
325
+ container_port: 12002
326
+ HOSPITAL_ENV_INJECTION_MCP_PORT:
327
+ default: 12005
328
+ container_port: 12005
329
+
330
+ finance:
331
+ docker_compose: "dt_arena/envs/finance/docker-compose.yml"
332
+ reset_endpoints:
333
+ api:
334
+ url: "http://127.0.0.1:${FINANCE_WEB_PORT}/api/v1/reset"
335
+ method: POST
336
+ ports:
337
+ FINANCE_WEB_PORT:
338
+ default: 5000
339
+ container_port: 5000
340
+
341
+ telegram:
342
+ docker_compose: "dt_arena/envs/telegram/docker-compose-hub.yml"
343
+ reset_endpoints:
344
+ api:
345
+ url: "http://127.0.0.1:${TELEGRAM_API_PORT}/api/v1/reset"
346
+ method: POST
347
+ ports:
348
+ TELEGRAM_PG_PORT:
349
+ default: 5472
350
+ container_port: 5432
351
+ TELEGRAM_API_PORT:
352
+ default: 8038
353
+ container_port: 8038
354
+
355
+ legal:
356
+ docker_compose: "dt_arena/envs/legal/docker-compose.yml"
357
+ reset_endpoints:
358
+ api:
359
+ url: "http://127.0.0.1:${LEGAL_WEB_PORT}/api/v1/reset"
360
+ method: POST
361
+ ports:
362
+ LEGAL_WEB_PORT:
363
+ default: 8003
364
+ container_port: 8003
365
+
366
+ macos:
367
+ docker_compose: "dt_arena/envs/macos/docker-compose.yml"
368
+ vm_profile: "dt_arena/envs/macos/vm.yaml" # ADL backend uses QemuVM
369
+ max_instances: 1
370
+ health_timeout: 300 # macOS VM needs ~3-5 minutes for first boot; loadvm ~30s
371
+ reset_endpoints:
372
+ api:
373
+ url: "http://127.0.0.1:${MCP_SERVICE_PORT}/reset"
374
+ method: POST
375
+ listeners:
376
+ url: "http://127.0.0.1:${EXFIL_PORT}/reset-all"
377
+ method: DELETE
378
+ reset_retries: 2
379
+ reset_retry_delay: 15
380
+ ports:
381
+ WEB_MANAGEMENT_PORT:
382
+ default: 8046
383
+ container_port: 8006
384
+ MCP_SERVICE_PORT:
385
+ default: 8044
386
+ container_port: 8005
387
+ EXFIL_PORT:
388
+ default: 9899
389
+ container_port: 9999
390
+ FTP_API_PORT:
391
+ default: 9898
392
+ container_port: 8080
393
+ DNS_API_PORT:
394
+ default: 9897
395
+ container_port: 8080
396
+
397
+ github:
398
+ docker_compose: "dt_arena/envs/github/docker/docker-compose-hub.yml"
399
+ reset_endpoints:
400
+ api:
401
+ url: "http://127.0.0.1:${GITHUB_API_PORT}/api/v1/reset"
402
+ method: POST
403
+ ports:
404
+ GITHUB_API_PORT:
405
+ default: 8045
406
+ container_port: 8083
407
+ GITHUB_UI_PORT:
408
+ default: 8046
409
+ container_port: 8045
410
+ GITHUB_PG_PORT:
411
+ default: 5455
412
+ container_port: 5432
413
+
414
+ gitlab:
415
+ docker_compose: "dt_arena/envs/gitlab/docker-compose-hub.yml"
416
+ disable_reuse: true
417
+ ports:
418
+ GITLAB_UI_PORT:
419
+ default: 8023
420
+ container_port: 80
421
+ GITLAB_SSH_PORT:
422
+ default: 2223
423
+ container_port: 22
424
+
425
+ customer_service:
426
+ docker_compose: "dt_arena/envs/customer_service/docker-compose.yml"
427
+ max_instances: 10
428
+ reset_endpoints:
429
+ api:
430
+ url: "http://127.0.0.1:${CUSTOMER_SERVICE_RESET_PORT}/reset"
431
+ method: POST
432
+ ports:
433
+ CUSTOMER_SERVICE_DB_PORT:
434
+ default: 5444
435
+ container_port: 5432
436
+ CUSTOMER_SERVICE_API_PORT:
437
+ default: 8080
438
+ container_port: 8080
439
+ CUSTOMER_SERVICE_RESET_PORT:
440
+ default: 8081
441
+ container_port: 8081
442
+
443
+ terminal:
444
+ docker_compose: "dt_arena/envs/terminal/docker-compose-hub.yml"
445
+ max_instances: 10
446
+ reset_script_timeout: 180
447
+ reset_scripts:
448
+ terminal-env: "/scripts/reset.sh"
449
+
450
+ research:
451
+ docker_compose: "dt_arena/envs/research/docker-compose-hub.yml"
452
+ reset_scripts:
453
+ research-env: "/reset.sh"
454
+
455
+ whatsapp:
456
+ docker_compose: "dt_arena/envs/whatsapp/docker-compose-hub.yml"
457
+ reset_endpoints:
458
+ api:
459
+ url: "http://127.0.0.1:${WHATSAPP_API_PORT}/api/v1/reset"
460
+ method: POST
461
+ ports:
462
+ WHATSAPP_PG_PORT:
463
+ default: 5473
464
+ container_port: 5473
465
+ WHATSAPP_API_PORT:
466
+ default: 8039
467
+ container_port: 8039
468
+ WHATSAPP_FRONTEND_PORT:
469
+ default: 8040
470
+ container_port: 80
471
+
472
+ googledocs:
473
+ docker_compose: "dt_arena/envs/googledocs/docker-compose-hub.yml"
474
+ reset_endpoints:
475
+ api:
476
+ url: "http://127.0.0.1:${GOOGLEDOCS_API_PORT}/api/v1/reset"
477
+ method: POST
478
+ ports:
479
+ GOOGLEDOCS_PG_PORT:
480
+ default: 5474
481
+ container_port: 5474
482
+ GOOGLEDOCS_API_PORT:
483
+ default: 8041
484
+ container_port: 8041
485
+ GOOGLEDOCS_FRONTEND_PORT:
486
+ default: 8042
487
+ container_port: 80
488
+
489
+ arxiv:
490
+ docker_compose: "dt_arena/envs/arxiv/docker-compose-hub.yml"
491
+ reset_endpoints:
492
+ api:
493
+ url: "http://127.0.0.1:${ARXIV_API_PORT}/api/v1/reset"
494
+ method: POST
495
+ reset_scripts:
496
+ arxiv-api: "/app/reset.sh"
497
+ ports:
498
+ ARXIV_API_PORT:
499
+ default: 8034
500
+ container_port: 8033
501
+ ARXIV_UI_PORT:
502
+ default: 8058
503
+ container_port: 80
504
+
505
+ bigquery:
506
+ docker_compose: "dt_arena/envs/bigquery/docker-compose.yml"
507
+ disable_reuse: true
508
+ health_timeout: 180
509
+ ports:
510
+ BIGQUERY_API_PORT:
511
+ default: 9050
512
+ container_port: 9050
513
+ BIGQUERY_GRPC_PORT:
514
+ default: 9060
515
+ container_port: 9060