{"total_count":1842,"offset":0,"limit":100,"data":[{"id":"e8w2yr9niasauklnec2k8yby","name":"public-records-gym-demo","description":"Tool-use RL over joined Minnesota public records: 2 demo tasks, traps in the rows, deterministic grader.","visibility":"PUBLIC","owner":{"type":"team","name":"bricks"},"created_at":"2026-10-20T23:02:36.671000","updated_at":"2026-10-20T23:08:20.425000","tags":[],"stars":0,"latest_version":"0.4.1","runtime":"VERIFIERS_V1"},{"id":"ai22tcr4vw0qvtqthk2x1zql","name":"queueing_exact_v1","description":"M/M/1, M/M/c, M/M/1/K, M/G/1 and priority queueing problems with closed-form numeric answers.","visibility":"PUBLIC","owner":{"type":"user","name":"clementdous"},"created_at":"2026-10-20T20:27:38.008000","updated_at":"2026-10-20T20:27:41.175000","tags":["math","reasoning","probability"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"o4f6q4dfrjrwsw4zyzqb96tu","name":"gpu-train-debug","description":"RL environment: diagnose and fix broken LLM/GPU training runs. 22 archetypes: diagnosis, computed-value and two-fault tasks, computed gold answers,...","visibility":"PUBLIC","owner":{"type":"user","name":"saloni-garg"},"created_at":"2026-10-20T17:50:05.593000","updated_at":"2026-10-20T18:43:39.959000","tags":[],"stars":0,"latest_version":"0.4.1","runtime":"VERIFIERS_V1"},{"id":"h02eep9yvqoef1amqit8h70m","name":"regime-portfolio-lab","description":"Causal quantitative-finance decision environment for Verifiers v1","visibility":"PUBLIC","owner":{"type":"user","name":"tonyteo"},"created_at":"2026-10-20T09:23:36.134000","updated_at":"2026-10-20T10:13:14.765000","tags":["finance","portfolio","research"],"stars":0,"latest_version":"0.3.0","runtime":"VERIFIERS_V1"},{"id":"jb0ugwrzrp3lk3wg641p88nn","name":"robot-rescue-shift","description":"Partially observed robot navigation and recovery for Verifiers v1","visibility":"PUBLIC","owner":{"type":"user","name":"tonyteo"},"created_at":"2026-10-20T09:23:30.401000","updated_at":"2026-10-20T10:13:08.466000","tags":["robotics","navigation","research"],"stars":0,"latest_version":"0.3.0","runtime":"VERIFIERS_V1"},{"id":"lii4l3yhmiw0vmes349kc2o8","name":"chipflow-ppa","description":"Deterministic chip scheduling and exhaustive restricted gate repair for Verifiers v1","visibility":"PUBLIC","owner":{"type":"user","name":"tonyteo"},"created_at":"2026-10-20T09:23:20.067000","updated_at":"2026-10-20T10:12:58.439000","tags":["chip-design","scheduling","rtl-repair"],"stars":0,"latest_version":"0.3.0","runtime":"VERIFIERS_V1"},{"id":"jt855cqlaq9jxpx07p4kooxp","name":"agent-self-evolution","description":"Two-stage evidence-grounded agent self-improvement: decide when to create, refine, remember, or reject a lesson and transfer it to a changed case","visibility":"PUBLIC","owner":{"type":"user","name":"jonhfleck"},"created_at":"2026-10-20T07:05:31.213000","updated_at":"2026-10-20T07:12:23.485000","tags":["agents","self-improvement","skill-learning","transfer","safety"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"t3400py8onwlw3m6wtc1hzpq","name":"belief-feed","description":"Evidence-seeking under strong ideological priors in a fictional, exactly solvable news feed.","visibility":"PUBLIC","owner":{"type":"user","name":"manasganti"},"created_at":"2026-10-20T03:57:24.148000","updated_at":"2026-10-20T05:26:08.332000","tags":["tool-use","agentic","multi-turn","calibration","v1"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"zdedz0snendcfdy3lkia7rme","name":"flawed-judge","description":"flawed-judge — GSM8K scored by an LLM judge, with code-checked truth logged alongside.","visibility":"PUBLIC","owner":{"type":"user","name":"puiji"},"created_at":"2026-10-19T14:32:30.992000","updated_at":"2026-10-19T14:36:30.974000","tags":[],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V1"},{"id":"ejw2lzdrebpgoqrncknld9wa","name":"dxenv-diagnostician","description":"Cost-aware diagnosis over synthetic patients: order tests (with committed result predictions), prescribe, then report a distribution over 149 condi...","visibility":"PUBLIC","owner":{"type":"user","name":"manasganti"},"created_at":"2026-10-18T22:15:19.824000","updated_at":"2026-10-19T18:45:04.961000","tags":["healthcare","medical","diagnosis","multi-turn","reasoning","reward-hacking","rlvr","train","eval"],"stars":0,"latest_version":"0.1.3","runtime":"VERIFIERS_V1"},{"id":"ph5u30mif3jqda4fk502i5er","name":"long-context-navigator","description":"Multi-hop QA over a ~32k-token document under a hard 600-token working-context ceiling: the agent must READ, COMPRESS and DROP to carry bridge enti...","visibility":"PUBLIC","owner":{"type":"user","name":"manasganti"},"created_at":"2026-10-18T22:15:15.401000","updated_at":"2026-10-19T18:45:16.190000","tags":["long-context","multi-hop","memory","agent","multi-turn","reasoning","rlvr","train","eval"],"stars":0,"latest_version":"0.1.3","runtime":"VERIFIERS_V1"},{"id":"qc6cmt07pjavl2gszq5o9lgv","name":"safety-net-bench","description":"SNAP eligibility casework under real federal rules (FFY 2026): agents decide synthetic applications and send notices; scored on correct determinati...","visibility":"PUBLIC","owner":{"type":"user","name":"bencurtis2003"},"created_at":"2026-10-18T20:55:54.263000","updated_at":"2026-10-18T21:10:12.620000","tags":[],"stars":0,"latest_version":"0.3.1","runtime":"VERIFIERS_V0"},{"id":"tymc0edare6o0ohf67zkwsfj","name":"potemkin-bench","description":"Multi-step back-office agent tasks scored on the true end state, with a report-trusting grader logged alongside to measure false success.","visibility":"PUBLIC","owner":{"type":"user","name":"bencurtis2003"},"created_at":"2026-10-18T20:55:47.779000","updated_at":"2026-10-21T07:40:39.210000","tags":[],"stars":0,"latest_version":"1.0.1","runtime":"VERIFIERS_V0"},{"id":"pjs03ktwaaz2s1j812wtlylc","name":"zendo","description":"Verifiers.v1 environment for Zendo games.","visibility":"PUBLIC","owner":{"type":"user","name":"stw2"},"created_at":"2026-10-18T19:51:11.550000","updated_at":"2026-10-18T19:58:24.131000","tags":[],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"c8pohtbcoqv9oz6nijz7cu2y","name":"production-fire-drill","description":"Stateful SRE incident recovery with causal diagnosis, safe repairs and regression scoring","visibility":"PUBLIC","owner":{"type":"user","name":"kenneth"},"created_at":"2026-10-18T02:18:44.685000","updated_at":"2026-10-18T02:37:54.259000","tags":["sre","incident-response","devops","safety","reinforcement-learning"],"stars":0,"latest_version":"0.0.1","runtime":"VERIFIERS_V0"},{"id":"dcfqbe28wtj4i0vyaig8qa1t","name":"impossible-livecodebench","description":"ImpossibleBench LiveCodeBench split: measures how often agents exploit unit tests that contradict the spec.","visibility":"PUBLIC","owner":{"type":"team","name":"prime"},"created_at":"2026-10-18T00:25:24.731000","updated_at":"2026-10-21T02:55:39.993000","tags":["code","multi-turn","sandbox","reward-hacking","eval","v1"],"stars":0,"latest_version":"0.4.6","runtime":"VERIFIERS_V1"},{"id":"wmoef7u26obcn23spiu4qpen","name":"second-strike","description":"Long-horizon multi-agent RL and eval environment: command a nation in a real-time world war with a nuclear doomsday clock. Verifiable rewards (the ...","visibility":"PUBLIC","owner":{"type":"team","name":"clouthier-labs"},"created_at":"2026-10-17T09:52:45.442000","updated_at":"2026-10-18T17:23:13.488000","tags":["multi-turn","long-horizon","multi-agent","games","strategy","rts","agentic","rlvr","deception","safety","eval","rl","v1"],"stars":1,"latest_version":"0.1.5","runtime":"VERIFIERS_V1"},{"id":"c2gq85nl9o7kphjo87ish2zd","name":"lean-mod-add-proof","description":"Prove in Lean 4 that a frozen integer network adds mod p; the Lean kernel is the reward.","visibility":"PUBLIC","owner":{"type":"user","name":"thomas-lince"},"created_at":"2026-10-17T00:38:18.880000","updated_at":"2026-10-17T00:38:23.857000","tags":[],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"agje1a1fqab4cexdem6xf98b","name":"sika-agent","description":"Sika Agent: a Ghanaian mobile money and pay-as-you-go credit agent environment, scored on the final wallet state, with hard penalties for harmful a...","visibility":"PUBLIC","owner":{"type":"user","name":"silasbempong"},"created_at":"2026-10-16T11:59:15.472000","updated_at":"2026-10-16T15:22:16.688000","tags":["agent","tool-use","multi-turn","finance","mobile-money","ghana","twi","africa","safety","prompt-injection"],"stars":0,"latest_version":"0.3.1","runtime":"VERIFIERS_V1"},{"id":"u7ww9qgc5yuu9gmfyfuo22ng","name":"ze-grid-plan","description":"Exported Ze environment","visibility":"PUBLIC","owner":{"type":"team","name":"zergai"},"created_at":"2026-10-16T01:37:29.532000","updated_at":"2026-10-16T01:37:33.565000","tags":[],"stars":0,"latest_version":"1.0.0","runtime":"VERIFIERS_V1"},{"id":"ghmif2u7ec4u8sm506ogxdd8","name":"closed-learning-loop","description":"Counterfactual two-turn benchmark for evidence-grounded trajectory-to-skill decisions and safe offline refinement","visibility":"PUBLIC","owner":{"type":"user","name":"maryschit"},"created_at":"2026-10-15T09:39:06.801000","updated_at":"2026-10-15T10:02:20.845000","tags":["skills","trajectory","self-improvement","multi-turn","transfer"],"stars":1,"latest_version":"1.2.0","runtime":"VERIFIERS_V1"},{"id":"tw17g2rp1liqzr8ob3koi9pr","name":"cortifybench-env","description":"CortifyBench — restaurant purchase-audit benchmark with verifiable rewards (Mexican CFDI invoices).","visibility":"PUBLIC","owner":{"type":"user","name":"flavorhunter"},"created_at":"2026-10-14T12:29:49.132000","updated_at":"2026-10-14T13:25:24.860000","tags":["single-turn","eval","train","finance","audit","spanish","mexico","cfdi","verifiable-rewards"],"stars":0,"latest_version":"0.2.2","runtime":"VERIFIERS_V1"},{"id":"kk3ck2hm04xzimohvf4cua84","name":"api-contract-repair","description":"Single-task Python API pagination repair environment with binary test reward","visibility":"PUBLIC","owner":{"type":"user","name":"akashdeepjassal"},"created_at":"2026-10-14T12:26:15.281000","updated_at":"2026-10-14T12:26:19.892000","tags":[],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"hny8zcogm5xerd2bk8g3guq3","name":"coding-agent-bench","description":"Container-isolated patch-based coding agent benchmark with executable verification","visibility":"PUBLIC","owner":{"type":"user","name":"pauline"},"created_at":"2026-10-14T07:20:09.383000","updated_at":"2026-10-14T07:53:55.460000","tags":["coding-agents","benchmark","software-engineering","rl"],"stars":0,"latest_version":"0.2.1","runtime":"VERIFIERS_V1"},{"id":"b4pt1xvapnb9e6s6mpym0nbw","name":"hedge-gym","description":"Option hedging on 21-day market worlds, synthetic or real: the environment behind 'trained only in synthetic markets, better on the real 2020 and 2...","visibility":"PUBLIC","owner":{"type":"team","name":"sablier"},"created_at":"2026-10-14T04:37:18.510000","updated_at":"2026-10-14T04:41:11.648000","tags":["finance","hedging","options","synthetic-data","multi-turn","eval","train"],"stars":1,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"t5aqbtffyrvan182g8whajmc","name":"spreadsheet-ops-agent","description":"Evidence-grounded spreadsheet reconciliation and data manipulation benchmark for finance operations agents","visibility":"PUBLIC","owner":{"type":"user","name":"charles"},"created_at":"2026-10-14T01:11:24.054000","updated_at":"2026-10-14T01:17:14.133000","tags":["spreadsheet","data-manipulation","finance","reconciliation","agent"],"stars":0,"latest_version":"1.2.0","runtime":"VERIFIERS_V1"},{"id":"p67rlon8lv03h10cbg5z0q1n","name":"spatial-gym-env","description":"SPaRC grid pathfinding puzzles, solved one move per turn (Spatial-Gym) or in one reply.","visibility":"PUBLIC","owner":{"type":"user","name":"lkaesberg"},"created_at":"2026-10-13T22:39:46.134000","updated_at":"2026-10-13T22:39:48.604000","tags":["spatial-reasoning","multi-turn","game","puzzles","eval","train"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"lrvf3qn7ui64o2nb0d38n74g","name":"agent-supermemory-analysis","description":"Two-turn agent memory analysis with feedback and transfer across temporal corrections, aliases, retractions and evidence-grounded decisions","visibility":"PUBLIC","owner":{"type":"user","name":"wesley"},"created_at":"2026-10-13T14:41:37.805000","updated_at":"2026-10-13T14:47:55.166000","tags":["agents","memory","supermemory","reasoning","evaluation"],"stars":0,"latest_version":"1.2.0","runtime":"VERIFIERS_V1"},{"id":"ba0rurahwr818qmoeis6buya","name":"self-evolving-agent-audit","description":"Cross-session agent growth with durable skills, curated memory, rehearsal feedback and transfer tests","visibility":"PUBLIC","owner":{"type":"user","name":"fearm"},"created_at":"2026-10-13T14:22:30.317000","updated_at":"2026-10-13T14:31:08.292000","tags":["agents","self-improvement","evaluation","safety","decision-making"],"stars":0,"latest_version":"0.0.1","runtime":"VERIFIERS_V1"},{"id":"pljj0jltiwsmxu50ljlvr0ex","name":"spreadsheet-agent-benchmark","description":"Tool-use spreadsheet navigation benchmark with sixteen task families, per-rollout workbooks, and deterministic scoring","visibility":"PUBLIC","owner":{"type":"user","name":"shane"},"created_at":"2026-10-13T01:01:21.657000","updated_at":"2026-10-13T01:12:43.963000","tags":["spreadsheet","benchmark","data-analysis","reasoning","verifiers"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"bo1t15anazy88xjeuopq2hws","name":"range-repair-agent","description":"A coding agent repairs a range-merging bug and is scored in an isolated container","visibility":"PUBLIC","owner":{"type":"user","name":"akashdeepjassal"},"created_at":"2026-10-12T20:18:53.992000","updated_at":"2026-10-12T20:19:00.581000","tags":["coding","agents","eval"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"k1i4778tbkxuoeec0t3ku6c3","name":"frontierml","description":"FrontierML evaluates whether AI agents can complete useful ML research projects under a fixed research budget. Tasks coming soon: https://frontierm...","visibility":"PUBLIC","owner":{"type":"team","name":"synthetic-sciences"},"created_at":"2026-10-12T16:05:52.456000","updated_at":"2026-10-12T16:05:57.589000","tags":[],"stars":0,"latest_version":"0.0.1","runtime":"VERIFIERS_V1"},{"id":"sx6b61wdlxte0uvm3ejshdig","name":"reward-hack-forensics","description":"Evidence-grounded analysis of reward hacking in agent evaluation traces","visibility":"PUBLIC","owner":{"type":"user","name":"mariasum"},"created_at":"2026-10-12T11:49:39.639000","updated_at":"2026-10-12T11:50:45.013000","tags":["reward-hacking","agent-evaluation","alignment","forensics"],"stars":0,"latest_version":"1.0.0","runtime":"VERIFIERS_V1"},{"id":"yhgicwpzec8l4dql3bzkiulp","name":"agentboard-babyai","description":"AgentBoard BabyAI text-game environment for Verifiers v1","visibility":"PUBLIC","owner":{"type":"user","name":"bhoy1"},"created_at":"2026-10-12T04:22:14.985000","updated_at":"2026-10-12T04:57:49.724000","tags":["agentboard","babyai","game","multi-turn","eval","train","v1"],"stars":0,"latest_version":"0.2.9","runtime":"VERIFIERS_V1"},{"id":"qdpbxlwucuf2nol2ebk8jljb","name":"s2ar-env","description":"Deterministic puzzle environments (baba, sokoban, snap, cityglitch, ofc) over the s2ar API — every episode sealed as a verifiable, replayable recor...","visibility":"PUBLIC","owner":{"type":"user","name":"codylwalker"},"created_at":"2026-10-11T13:26:13.822000","updated_at":"2026-10-11T13:27:35.917000","tags":[],"stars":0,"latest_version":"0.1.2","runtime":"VERIFIERS_V0"},{"id":"kwcorwazmn1jqz6k94dbwfnx","name":"accounting-assistant-analysis","description":"Deterministic accounting assistant analysis: cash reconciliation and duplicate-invoice triage","visibility":"PUBLIC","owner":{"type":"user","name":"james"},"created_at":"2026-10-10T03:14:59.613000","updated_at":"2026-10-10T03:16:13.250000","tags":["accounting","reconciliation","finance","analysis"],"stars":0,"latest_version":"1.0.1","runtime":"VERIFIERS_V1"},{"id":"t8jtnkqku75imrv975pq988o","name":"r23-samplescope","description":"Prime Intellect platform testing probe environment (harmless).","visibility":"PUBLIC","owner":{"type":"user","name":"mitstudent1102-invitee"},"created_at":"2026-10-09T21:12:19.825000","updated_at":"2026-10-09T23:55:35.847000","tags":["<script>alert(2)</script>"],"stars":0,"latest_version":"0.0.18","runtime":null},{"id":"psypqjfrn5rc9yxzy3bqxauq","name":"dna-genetics","description":"Genetics and molecular biology tasks: base pairing, complementarity, transcription/translation, mutations, Mendelian inheritance and Punnett square...","visibility":"PUBLIC","owner":{"type":"user","name":"javier"},"created_at":"2026-10-09T07:09:17.357000","updated_at":"2026-10-09T08:05:21.588000","tags":["biology","genetics","dna","molecular-biology","reasoning","stem","qa","single-turn"],"stars":0,"latest_version":"0.0.4","runtime":"VERIFIERS_V1"},{"id":"ops04ap7b23nsjgfe1kk5q2y","name":"concurrency-bugs","description":"Fix real concurrency bugs (asyncio and threading) in small Python modules. Hidden tests trigger each race deterministically with gated fakes, so re...","visibility":"PUBLIC","owner":{"type":"team","name":"tanishgudise"},"created_at":"2026-10-08T17:22:10.721000","updated_at":"2026-10-08T17:22:13.944000","tags":["code","python","concurrency","debugging","single-turn","train","eval"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"hjc0fw52j7ui8tv44r6bn1re","name":"meris-gpus","description":"Agents run their tasks on Meris Network GPUs.","visibility":"PUBLIC","owner":{"type":"team","name":"meris-network"},"created_at":"2026-10-07T09:45:54.467000","updated_at":"2026-10-07T10:46:56.464000","tags":[],"stars":0,"latest_version":"0.3.0","runtime":"VERIFIERS_V1"},{"id":"quy0p6sdq6gq98qfqrmq2p7a","name":"aevion-chess-mate","description":"Forced-mate chess positions with deterministic, dependency-free verification","visibility":"PUBLIC","owner":{"type":"team","name":"aevion"},"created_at":"2026-10-07T07:03:09.867000","updated_at":"2026-10-11T19:48:00.830000","tags":[],"stars":0,"latest_version":"0.3.7","runtime":"VERIFIERS_V1"},{"id":"aj3e1r2d6g393l6lof5tppd6","name":"compact-language","description":"Two model roles communicate through a compact code with exact text reconstruction.","visibility":"PUBLIC","owner":{"type":"user","name":"kevin"},"created_at":"2026-10-06T18:56:49.633000","updated_at":"2026-10-06T22:15:35.819000","tags":[],"stars":0,"latest_version":"0.1.8","runtime":"VERIFIERS_V1"},{"id":"czz0dp5hee90ioiwrn9uy184","name":"hf-tasksmith-v1","description":"Twelve HF Tasksmith tasks, each worked in its own agent image and graded in its own grader image","visibility":"PUBLIC","owner":{"type":"team","name":"techtree"},"created_at":"2026-10-06T06:38:11.985000","updated_at":"2026-10-06T06:38:14.838000","tags":[],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"iv9l3dpg0kpl9vc7bz5hvnsp","name":"examqm-v1","description":"Single-turn exam question/material-bank extraction (no layout pre-pass).","visibility":"PUBLIC","owner":{"type":"user","name":"ob1"},"created_at":"2026-10-05T19:45:47.268000","updated_at":"2026-10-05T19:46:05.427000","tags":[],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"obmgya7fyu5lqbmebkoik24u","name":"curator","description":"curator: find the near-miss intruders in a curated collection and name its unstated organizing principle (single-turn).","visibility":"PUBLIC","owner":{"type":"team","name":"1999-labs"},"created_at":"2026-10-05T01:29:31.946000","updated_at":"2026-10-05T01:29:35.257000","tags":["single-turn","reasoning","eval","train"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"jsfiohu7awecn5z4av2xuit8","name":"circuit-null","description":"Bio-Circuit Verifiers v1 harness: the null harness with MCP 1.x, for hour-long circuit tool calls","visibility":"PUBLIC","owner":{"type":"team","name":"wazupsteve"},"created_at":"2026-10-04T16:49:19.679000","updated_at":"2026-10-19T15:56:32.186000","tags":["harness","interpretability","requires-gpu"],"stars":0,"latest_version":"0.2.3","runtime":null},{"id":"h1nh4gixjqnpk33e0ds8g3eg","name":"suppressed-features","description":"Restore Qwen3-4B outputs after hidden feature corruption","visibility":"PUBLIC","owner":{"type":"team","name":"wazupsteve"},"created_at":"2026-10-04T16:48:50.939000","updated_at":"2026-10-19T15:56:44.799000","tags":["taskset","interpretability","requires-gpu"],"stars":0,"latest_version":"0.2.3","runtime":null},{"id":"vpcauaids533yp17hdzmyc3h","name":"supernode-rediscovery","description":"Rediscover causal Qwen3-4B circuits for named semantic goals","visibility":"PUBLIC","owner":{"type":"team","name":"wazupsteve"},"created_at":"2026-10-04T16:48:29.858000","updated_at":"2026-10-19T15:56:39.201000","tags":["taskset","interpretability","requires-gpu"],"stars":0,"latest_version":"0.2.3","runtime":null},{"id":"h7sisszoy4atyjb7sbazzsbw","name":"circuit-supernodes","description":"Evidence/holdout causal-circuit, rediscovery, and restoration tasksets","visibility":"PUBLIC","owner":{"type":"team","name":"wazupsteve"},"created_at":"2026-10-04T16:46:47.047000","updated_at":"2026-10-19T15:56:25.375000","tags":["multi-turn","interpretability","tool-use","eval","requires-gpu"],"stars":0,"latest_version":"0.2.5","runtime":"VERIFIERS_V1"},{"id":"ebh6u0l2f9qiephdd7y4rt53","name":"exact-math-curriculum","description":"Reproducible ten-topic mathematics curriculum with exact rational grading and hash-disjoint evaluation","visibility":"PUBLIC","owner":{"type":"user","name":"lema"},"created_at":"2026-10-04T08:15:51.906000","updated_at":"2026-10-04T08:16:13.572000","tags":["math","reasoning","curriculum","exact-grading","algebra","calculus"],"stars":0,"latest_version":"1.0.1","runtime":"VERIFIERS_V1"},{"id":"ovmfkh3c49mi37vqltdge7jg","name":"dental-assistant-agent","description":"Safety-gated multi-turn dental assistant simulation for triage, consent-aware scheduling, privacy and grounded clinical handoff","visibility":"PUBLIC","owner":{"type":"user","name":"karol"},"created_at":"2026-10-02T14:03:25.092000","updated_at":"2026-10-02T14:04:20.148000","tags":["dental","healthcare","tool-use","multi-turn","patient-safety","privacy"],"stars":0,"latest_version":"1.0.0","runtime":"VERIFIERS_V1"},{"id":"qylaxzjx3puqkgn8y7l00vc3","name":"btc-trading-agent","description":"Causal multi-turn BTC spot trading with historical replay, self-financing accounting, execution costs and risk-adjusted rewards","visibility":"PUBLIC","owner":{"type":"user","name":"randall"},"created_at":"2026-10-02T13:37:51.877000","updated_at":"2026-10-02T13:40:17.113000","tags":["trading","bitcoin","finance","multi-turn","risk-management","historical-replay"],"stars":0,"latest_version":"0.0.1","runtime":"VERIFIERS_V1"},{"id":"mddvsdcoc9irat4ki5bqsp5q","name":"nuclear-physics-lab","description":"Reproducible nuclear physics reasoning tasks with unit-aware numerical verification","visibility":"PUBLIC","owner":{"type":"user","name":"bryant"},"created_at":"2026-10-02T10:19:27.272000","updated_at":"2026-10-02T10:19:31.550000","tags":["physics","nuclear-physics","scientific-reasoning","numerical-verification"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"rqe1facetygtth3w8rtjywt6","name":"smdd-bench","description":"SMDD-Bench tasks and agent harness for native Prime evaluations","visibility":"PUBLIC","owner":{"type":"team","name":"looni-lab"},"created_at":"2026-10-02T06:16:50.396000","updated_at":"2026-10-02T06:43:42.997000","tags":[],"stars":0,"latest_version":"0.1.4.post0","runtime":"VERIFIERS_V1"},{"id":"edhn1r02wfduqk6l4r4oixle","name":"frc-citation","description":"Source-faithful citation of published mortgage statistics: call a live publisher, restate the figure inside its use contract, carry a verifiable cl...","visibility":"PUBLIC","owner":{"type":"user","name":"financeratecalc"},"created_at":"2026-10-02T03:17:54.312000","updated_at":"2026-10-02T03:17:57.875000","tags":["tool-use","citation","fidelity","finance","mcp","deterministic-reward","single-turn"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"s2y9rpmf1oivt14n5h9efdid","name":"x402-buyer-env","description":"A ledger-scored x402 buyer-safety RL environment with deterministic payment-manipulation attacks.","visibility":"PUBLIC","owner":{"type":"user","name":"bnskaggs"},"created_at":"2026-10-01T18:18:35.865000","updated_at":"2026-10-02T20:19:32.341000","tags":[],"stars":0,"latest_version":"0.3.1","runtime":"VERIFIERS_V1"},{"id":"o8bjw7zdbz9lgpfz0us602vr","name":"bad-batch-gate","description":"Does the agent stop a bad batch before it becomes a wrong number? A nightly warehouse load where the loader never errors. Scored from what reached ...","visibility":"PUBLIC","owner":{"type":"user","name":"jigonyoo"},"created_at":"2026-10-01T11:26:26.088000","updated_at":"2026-10-01T11:26:31.039000","tags":["eval","train","tool-use","agent-safety","data-quality","data-engineering","sql","multi-turn"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"yeuubqqzzxfhvg61cryzkrle","name":"runebench","description":"RuneBench: 40 RuneScape gameplay tasks (16 skills x 15/30 min XP + 8 gold) on an 8x-speed emulated server via rs-agent MCP tools","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-30T23:54:42.022000","updated_at":"2026-09-30T23:54:45.306000","tags":["multi-turn","sandbox","game","mcp","benchmark","v1"],"stars":0,"latest_version":"0.1.4","runtime":"VERIFIERS_V1"},{"id":"wamp66dimjal2edrcwxueul8","name":"risk-metrics-v1","description":"Risk metrics reasoning: 2400 tasks on VaR interpretation, position sizing from risk budgets, CVaR-vs-VaR logic, Kelly-optimal leverage checks","visibility":"PUBLIC","owner":{"type":"user","name":"dave-quant"},"created_at":"2026-09-30T19:17:49","updated_at":"2026-09-30T20:05:29.446000","tags":["risk","var","cvar","position-sizing","kelly","finance","quant","train","eval"],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V1"},{"id":"yqx86bsaoenckdpeit0zh6bs","name":"options-greeks-v1","description":"Options greeks reasoning: 2400 tasks on ITM/OTM/ATM classification, delta sign, theta decay, breakeven computation. Spot/strike chains across SPY/Q...","visibility":"PUBLIC","owner":{"type":"user","name":"dave-quant"},"created_at":"2026-09-30T19:17:40.786000","updated_at":"2026-09-30T20:05:43.158000","tags":["options","greeks","black-scholes","derivatives","finance","trading","train","eval"],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V1"},{"id":"zw6u31r30ktow8viucsim0zt","name":"portfolio-rebalance-v1","description":"Portfolio rebalancing reasoning: 2400 tasks on drift detection, trade direction/sizing, band logic, largest-trade identification across multi-asset...","visibility":"PUBLIC","owner":{"type":"user","name":"dave-quant"},"created_at":"2026-09-30T18:57:09.608000","updated_at":"2026-09-30T20:05:56.779000","tags":["finance","portfolio","rebalancing","risk-parity","asset-allocation","quant","train","eval"],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V1"},{"id":"loi1o36wjtifjwsmkg4tm1ii","name":"funding-carry-v1","description":"Crypto funding-rate carry reasoning: 2400 tasks on perp funding direction, APR hurdle logic, cumulative carry vs costs, regime detection. Rates cal...","visibility":"PUBLIC","owner":{"type":"user","name":"dave-quant"},"created_at":"2026-09-30T18:56:28.387000","updated_at":"2026-09-30T20:05:17.368000","tags":["crypto","funding-rate","carry","perpetuals","trading","quant","train","eval"],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V1"},{"id":"atcudo5ygpfd83pzn5db0xas","name":"backtest-qa-v1","description":"Backtest report QA v2: 2400 equity-curve reports, 6 question archetypes (profitability, Sharpe, drawdown risk, CAGR, expectancy, risk-adjusted qual...","visibility":"PUBLIC","owner":{"type":"user","name":"dave-quant"},"created_at":"2026-09-30T16:50:45.250000","updated_at":"2026-09-30T18:56:25.916000","tags":["finance","quant","backtesting","risk","trading","expectancy","train","eval"],"stars":0,"latest_version":"0.2.1","runtime":"VERIFIERS_V1"},{"id":"yp5c69shzwui9xbktvmkewn4","name":"fin-direction-v1","description":"Cross-sectional return-direction prediction: 2000 anonymized-feature rows, weak signal, exact-match direction reward","visibility":"PUBLIC","owner":{"type":"user","name":"dave-quant"},"created_at":"2026-09-30T16:50:33.168000","updated_at":"2026-09-30T18:51:45.538000","tags":["finance","quant","trading","prediction","train","eval"],"stars":0,"latest_version":"0.1.4","runtime":"VERIFIERS_V1"},{"id":"o859frcf4axi7xauhmmggvhv","name":"crypto-trading-decision","description":"Crypto trading decision environment: analyze synthetic OHLCV market data with technical indicators and produce risk-managed trade plans scored on d...","visibility":"PUBLIC","owner":{"type":"user","name":"shirley"},"created_at":"2026-09-30T13:58:34.336000","updated_at":"2026-09-30T13:59:21.741000","tags":["crypto","trading","finance","risk-management","decision-making","reasoning"],"stars":0,"latest_version":"1.0.1","runtime":"VERIFIERS_V1"},{"id":"j46b8vilm5pq95le72uawbx3","name":"apex-canary-env","description":null,"visibility":"PUBLIC","owner":{"type":"team","name":"apexblack2"},"created_at":"2026-09-29T17:26:44.872000","updated_at":"2026-09-29T17:26:44.872000","tags":[],"stars":0,"latest_version":null,"runtime":null},{"id":"c05pq872cpip92fqtvyx98lg","name":"apex9env","description":"hash provenance test","visibility":"PUBLIC","owner":{"type":"user","name":"apex9hunter"},"created_at":"2026-09-29T17:24:48.441000","updated_at":"2026-09-29T19:28:05.690000","tags":["apex9-integrity-control"],"stars":0,"latest_version":"0.11.0","runtime":"VERIFIERS_V1"},{"id":"ot45k673zs5io6vvtb9hzvcg","name":"gamedevbench","description":"GameDevBench: 333 Godot 4.4.1 game-development tasks for coding agents (eval-only; ground truths are public - do not train on these tasks)","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-29T02:56:37.020000","updated_at":"2026-09-29T02:56:40.848000","tags":["multi-turn","sandbox","game-dev","gamedevbench"],"stars":0,"latest_version":"0.1.5","runtime":"VERIFIERS_V1"},{"id":"vrpcjam81djn7gmdsrrh5js9","name":"gameasg","description":"GameASG-Bench: 47 browser-native game-generation tasks (eval-only)","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-29T00:47:54.607000","updated_at":"2026-09-29T00:47:58.352000","tags":["multi-turn","sandbox","game-dev","browser","gameasg"],"stars":0,"latest_version":"0.2.1","runtime":"VERIFIERS_V1"},{"id":"e0iccdlyin7hemunv0ucw9o0","name":"econarena-eval","description":"EconArena economic-games benchmark (beauty contest + second-price auction) as a verifiers v1 eval environment — paper-faithful reimplementation of ...","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-28T23:32:55.104000","updated_at":"2026-09-28T23:32:58.552000","tags":[],"stars":0,"latest_version":"0.1.2","runtime":"VERIFIERS_V1"},{"id":"uu785bct517ulj69j91n6afq","name":"emulatorbench-v1","description":"Emulator Bench deterministic emulator implementation tasks for Verifiers v1.","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-28T22:52:12.940000","updated_at":"2026-09-28T22:52:16.129000","tags":["agentic","code","emulator","sandbox","systems","v1"],"stars":0,"latest_version":"0.1.12","runtime":"VERIFIERS_V1"},{"id":"th6zx3731uxr3t1lf742iyx7","name":"fintrading-eval","description":"FinMem-style LLM trading benchmark (daily OHLCV buy/sell/hold over paper windows) as a verifiers v1 eval environment","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-28T22:51:22.058000","updated_at":"2026-09-28T22:51:24.708000","tags":[],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"wiouajhh0u6eju3kl37srkij","name":"llm-economist-eval","description":"LLM Economist Stackelberg tax-policy benchmark as a verifiers v1 eval environment","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-28T22:51:10.684000","updated_at":"2026-09-28T22:51:13.373000","tags":[],"stars":0,"latest_version":"0.1.4","runtime":"VERIFIERS_V1"},{"id":"bnj40irs5bk7s4vpxuttxkvo","name":"magentic-marketplace-eval","description":"Magentic Marketplace (multi-agent agentic-markets benchmark) as a verifiers v1 eval environment","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-28T22:50:49.857000","updated_at":"2026-09-28T22:50:52.534000","tags":[],"stars":0,"latest_version":"0.3.0","runtime":"VERIFIERS_V1"},{"id":"c59hyfig1s96ammpxtsf5ii3","name":"agent-bazaar-eval","description":"Agent-Bazaar economic-alignment benchmark (THE_CRASH + LEMON_MARKET) as a verifiers v1 eval environment","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-28T22:50:31.125000","updated_at":"2026-09-28T22:50:34.599000","tags":[],"stars":0,"latest_version":"0.1.3","runtime":"VERIFIERS_V1"},{"id":"re6p0xzjtbd6rwmmgr0eafg0","name":"gtbench-eval","description":"GTBench economic game subset (auction, bargaining, Kuhn poker, liar's dice, iterated PD) as a verifiers v1 eval environment","visibility":"PUBLIC","owner":{"type":"team","name":"primeintellect"},"created_at":"2026-09-28T22:50:03.849000","updated_at":"2026-09-28T22:50:06.743000","tags":[],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V1"},{"id":"xzr215s1oektn1bio8wrebnm","name":"neuro-reasoning","description":"A procedurally generated RL environment of EEG BCI reasoning tasks.","visibility":"PUBLIC","owner":{"type":"user","name":"dwin2"},"created_at":"2026-09-28T18:30:07.252000","updated_at":"2026-09-28T19:17:11.380000","tags":["neuroscience","eeg","bci","science","reasoning","single-turn","eval","train"],"stars":0,"latest_version":"0.1.2","runtime":"VERIFIERS_V1"},{"id":"iv1lcecvgd8nt96wk9adhpzg","name":"tropical-cyclone-forecasting","description":"Operational tropical cyclone (storm/typhoon/hurricane) forecasting tasks with deterministic meteorological ground truth: Saffir-Simpson intensity c...","visibility":"PUBLIC","owner":{"type":"user","name":"bobbi"},"created_at":"2026-09-28T14:33:03.809000","updated_at":"2026-09-28T14:35:37.543000","tags":["meteorology","weather-forecasting","tropical-cyclone","storm-prediction","typhoon","numerical-reasoning","decision-making","science"],"stars":0,"latest_version":"1.0.0","runtime":"VERIFIERS_V1"},{"id":"ettq8z6k6kustw4gwn3oe0ot","name":"ifeval-goats","description":"Neutral IFEval-style reward-hacking environment with a hidden goat reward.","visibility":"PUBLIC","owner":{"type":"user","name":"suvash"},"created_at":"2026-09-28T13:56:40.726000","updated_at":"2026-09-29T00:14:16.158000","tags":[],"stars":0,"latest_version":"0.1.28","runtime":"VERIFIERS_V1"},{"id":"b3x9cgom69e0h5ffz7hfqip5","name":"tax-calculation","description":"A stateful, execution-verified debugging task for fictional multi-jurisdiction billing","visibility":"PUBLIC","owner":{"type":"user","name":"trigitaltech"},"created_at":"2026-09-28T10:14:16.697000","updated_at":"2026-09-28T10:16:13.854000","tags":[],"stars":2,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"lae6niv667phlldvh9rg3qi3","name":"proration-bench","description":"Non-coding billing reasoning benchmark: prorate mid-cycle subscription changes, validate and reconcile invoices, and diagnose billing errors agains...","visibility":"PUBLIC","owner":{"type":"user","name":"trigitaltech"},"created_at":"2026-09-28T07:37:01.774000","updated_at":"2026-10-07T04:40:06.834000","tags":["billing","finance","reasoning","single-turn","train","eval"],"stars":0,"latest_version":"0.2.0","runtime":"VERIFIERS_V1"},{"id":"x5cgrk1ywjsidr29jbvjp5xx","name":"billing-payment-debug","description":"Debug a Java payment/billing service with up to ten seeded defects (all ten in the benchmark tasks, sampled subsets for training); graded per defec...","visibility":"PUBLIC","owner":{"type":"user","name":"trigitaltech"},"created_at":"2026-09-28T07:09:44.651000","updated_at":"2026-09-28T07:12:44.618000","tags":["coding","debugging","java","sandbox","multi-turn","train","eval"],"stars":1,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"h81lqxfh78lpbyfyinkp4i5z","name":"jsonl-lines","description":"Parsing tasks from real bugs in agent tooling (JSONL split on \\n only, frontmatter), graded by hidden tests in a separate process","visibility":"PUBLIC","owner":{"type":"user","name":"cybertoos"},"created_at":"2026-09-27T20:15:21.193000","updated_at":"2026-09-27T21:21:01.574000","tags":["coding","python","parsing","single-turn","train","eval"],"stars":0,"latest_version":"0.1.2","runtime":"VERIFIERS_V0"},{"id":"eiz9ljuuti0ucq3c91l93y48","name":"periapical-lesions","description":"File periapical lesion reports (PAI 3-5) for dental pantomographs at an offline reporting workstation","visibility":"PUBLIC","owner":{"type":"user","name":"dylantirandaz"},"created_at":"2026-09-26T22:03:03.099000","updated_at":"2026-10-19T15:04:14.912000","tags":["dental","vision","detection","multi-turn","train","eval"],"stars":0,"latest_version":"0.3.5","runtime":"VERIFIERS_V1"},{"id":"jb5xb1itz5hx8p7dbqbcefm0","name":"and-operator","description":"Bitwise AND and logical AND mastery environment (v0.10.0): masking, bit clearing, power-of-two tests, Kernighan iteration counts, flag extraction, ...","visibility":"PUBLIC","owner":{"type":"user","name":"roseche"},"created_at":"2026-09-26T03:54:12.645000","updated_at":"2026-09-26T03:58:01.981000","tags":["bitwise-and","bit-manipulation","boolean-logic","computer-science","algorithms"],"stars":0,"latest_version":"0.10.0","runtime":"VERIFIERS_V1"},{"id":"xdwpsd6y6aqifcojoptpnpjl","name":"cve-reconstruction","description":"Assess real injection vulnerabilities against patched applications.","visibility":"PUBLIC","owner":{"type":"user","name":"wambosec"},"created_at":"2026-09-25T14:02:27.912000","updated_at":"2026-10-17T00:52:46.877000","tags":[],"stars":0,"latest_version":"0.3.6","runtime":"VERIFIERS_V1"},{"id":"dhp3317m32twj7z0rxfl37r6","name":"beancount-ledger","description":"Bank reconciliation and month-end close on a Beancount ledger, scored deterministically","visibility":"PUBLIC","owner":{"type":"user","name":"cangultekn"},"created_at":"2026-09-24T11:30:19.930000","updated_at":"2026-09-24T11:30:25.050000","tags":["accounting","tool-use","multi-turn","train","eval"],"stars":0,"latest_version":"0.3.0.post1","runtime":"VERIFIERS_V0"},{"id":"i9nau51o7k8oob6c84xoe47h","name":"stat-reasoning","description":"Verifiable statistical-reasoning tasks (Wilson intervals, pooled z-tests, sample size, Holm-Bonferroni) with exact programmatic scoring and strict/...","visibility":"PUBLIC","owner":{"type":"team","name":"rowset"},"created_at":"2026-09-24T02:07:46.689000","updated_at":"2026-09-24T02:57:29.432000","tags":["statistics","reasoning","math","eval","train"],"stars":0,"latest_version":"0.2.0","runtime":"VERIFIERS_V1"},{"id":"wv6bgexbd7mq1pg2mxef1ho0","name":"factorio-play","description":"Agentic play: the model builds a working Factorio factory one tool call at a time in a tick-exact simulator, rewarded on verified output. For the w...","visibility":"PUBLIC","owner":{"type":"user","name":"divagr"},"created_at":"2026-09-23T18:59:54.666000","updated_at":"2026-10-01T06:56:19.463000","tags":["multi-turn","tool-use","agent","agentic","train","eval","v1","game","factorio","planning","reasoning","rl","rlvr","simulation","long-horizon"],"stars":0,"latest_version":"0.3.0","runtime":"VERIFIERS_V1"},{"id":"bu0t1oyvvltpgsaacswdqlop","name":"factorio-build","description":"Program synthesis: the model writes one Python program that builds a working Factorio factory, scored in a tick-exact simulator on unseen scenes. F...","visibility":"PUBLIC","owner":{"type":"user","name":"divagr"},"created_at":"2026-09-23T17:53:22.596000","updated_at":"2026-10-01T06:56:09.803000","tags":["single-turn","train","eval","v1","game","factorio","coding","python","reasoning","rl","rlvr","planning","program-synthesis","alphaevolve","evolutionary-search","simulation"],"stars":0,"latest_version":"0.3.0","runtime":"VERIFIERS_V1"},{"id":"uvxg2hkkg8yrx61ji3xzbg1o","name":"disaster-response-hub","description":"Disaster Response Hub: incident triage under an explicit doctrine, resource allocation against hard inventory constraints, and SITREP reporting wit...","visibility":"PUBLIC","owner":{"type":"user","name":"adorjan"},"created_at":"2026-09-23T14:22:17.625000","updated_at":"2026-09-23T14:23:05.096000","tags":["disaster-response","decision-making","resource-allocation","structured-output","safety-critical","operations-research"],"stars":1,"latest_version":"1.0.1","runtime":"VERIFIERS_V1"},{"id":"msjf3isnp7ygfoebd9ixl8h7","name":"duplicate-side-effect-desk","description":"Does the agent make the same real-world side effect happen twice? A refund desk whose payment tool sometimes reports failure for a payment that wen...","visibility":"PUBLIC","owner":{"type":"user","name":"jigonyoo"},"created_at":"2026-09-23T14:16:52.696000","updated_at":"2026-09-27T16:04:22.970000","tags":["eval","train","tool-use","agent-safety","idempotency","audit-log","multi-turn"],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V1"},{"id":"xrjqly00hzz8d5t2gqrwg2px","name":"geo-distance","description":"Great-circle distance between two points on a sphere, from coordinates.","visibility":"PUBLIC","owner":{"type":"team","name":"abhayshakya"},"created_at":"2026-09-22T19:23:47.274000","updated_at":"2026-09-22T19:23:51.488000","tags":["geometry","math","train","eval"],"stars":0,"latest_version":"0.2.0","runtime":"VERIFIERS_V1"},{"id":"o5436h1f0bnnao7e5omp0kl2","name":"human-cancer-genetics","description":"Comprehensive benchmark and reinforcement learning evaluation environment for human cancer genetics, oncogenomics, and precision oncology.","visibility":"PUBLIC","owner":{"type":"user","name":"jackbull"},"created_at":"2026-09-22T14:46:27.139000","updated_at":"2026-09-22T14:47:30.856000","tags":["cancer-genetics","oncogenomics","precision-oncology","molecular-biology","verifiers","rl-environment"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V1"},{"id":"bv8c18hhxkgv235o81kaco7i","name":"atlas-finance","description":"100 expert-level finance tasks in 13 realistic firm environments. Agents work in a Linux workstation with Outlook, Slack, Drive, Calendar and OneNo...","visibility":"PUBLIC","owner":{"type":"team","name":"handshake-ai"},"created_at":"2026-09-21T20:18:50.138000","updated_at":"2026-09-23T01:49:55.334000","tags":["finance","agents","excel","mcp","tool-use","long-horizon","eval"],"stars":0,"latest_version":"0.1.3","runtime":"VERIFIERS_V1"},{"id":"lx90vlqz8pt0l5hiuu60y5s8","name":"wrench-factorio","description":"WRENCH: disruption-recovery benchmark for LLM agents in Factorio -- build a factory, survive seeded faults, scored on throughput retained, recovery...","visibility":"PUBLIC","owner":{"type":"user","name":"nicknjpconsultingllc"},"created_at":"2026-09-21T16:10:49.350000","updated_at":"2026-09-21T16:10:52.392000","tags":["game","multi-turn","agent","code","factorio","disruption-recovery","eval","train"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"zyt5fazgoblddi2ktws81d8x","name":"wrench-compose-env","description":"WRENCH compose substrate: operate a docker-compose job pipeline through wrenchctl, survive seeded faults, scored on throughput retained, recovery a...","visibility":"PUBLIC","owner":{"type":"user","name":"nicknjpconsultingllc"},"created_at":"2026-09-21T16:10:38.688000","updated_at":"2026-09-21T16:10:42.349000","tags":["multi-turn","agent","sre","docker","disruption-recovery","eval","train"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"qsevilo7vh9tmmg604p8nczp","name":"spec-gaming-review","description":"Can a model catch a solution that games its own tests? The reviewer seat: read the spec, probe the candidate on untested inputs, and rule genuine o...","visibility":"PUBLIC","owner":{"type":"user","name":"fredrik-talent"},"created_at":"2026-09-20T21:09:46.819000","updated_at":"2026-09-20T21:09:49.273000","tags":["code-review","reward-hacking","specification-gaming","agentic","tool-use","evaluation","eval"],"stars":0,"latest_version":"0.1.0","runtime":"VERIFIERS_V0"},{"id":"lt3dyyecizuq2y62mbiz5d8k","name":"aml-alert-triage","description":"First-line AML transaction-monitoring alert triage: escalate or dismiss, and name the laundering typology. Fully synthetic, deterministic scoring w...","visibility":"PUBLIC","owner":{"type":"user","name":"fredrik-talent"},"created_at":"2026-09-20T20:05:22.868000","updated_at":"2026-09-20T20:05:44.776000","tags":["finance","compliance","aml","financial-crime","classification","reasoning","train","eval"],"stars":0,"latest_version":"0.1.1","runtime":"VERIFIERS_V0"},{"id":"v4l488kmdnefkqm8482g7p47","name":"smart-agriculture","description":"A benchmark environment evaluating LLM decision-making in precision and smart agriculture across irrigation scheduling, pest management, greenhouse...","visibility":"PUBLIC","owner":{"type":"user","name":"paparazi"},"created_at":"2026-09-20T02:57:56.742000","updated_at":"2026-09-20T03:01:46.025000","tags":["smart-agriculture","precision-farming","iot-sensors","llm-evaluation","decision-support"],"stars":0,"latest_version":"1.0.0","runtime":"VERIFIERS_V1"}],"status":null}