| #799 — fix: preserve optional fields in structured_outputs strictify_schema_json |
@jkyi-nvidia |
review requested |
2026-03-02 20:59 UTC |
2026-03-03 20:59 UTC |
139d 16h |
| #879 — Yev/scicode env v1 |
@dhruvnathawani, @jubick1337 |
review requested |
2026-03-20 18:09 UTC |
2026-03-23 18:09 UTC |
125d 19h |
| #946 — Turing/ Playwright-based CUA environment with multi-provider adapter support |
@bxyu-nvidia |
review requested |
2026-04-29 19:09 UTC |
2026-04-30 19:09 UTC |
97d 18h |
| #1193 — docs: address repo-health audit nits in README and CONTRIBUTING |
@cmunley1 |
review requested |
2026-04-30 13:30 UTC |
2026-05-01 13:30 UTC |
96d 23h |
| #1264 — feat: artifact-root resolution for external plugins |
@ananthsub, @bxyu-nvidia |
review requested |
2026-05-07 20:20 UTC |
2026-05-08 20:20 UTC |
91d 17h |
| #1334 — structured_outputs: add LLM-as-a-judge semantic verification |
@jkyi-nvidia |
review requested |
2026-05-18 20:37 UTC |
2026-05-19 20:37 UTC |
84d 16h |
| #1439 — perf(responses_api_agents): bypass HTTP self-calls in all agents |
@bxyu-nvidia, @cmunley1, @ffrujeri |
review requested |
2026-05-28 07:19 UTC |
2026-05-29 07:19 UTC |
77d 6h |
| #1491 — Add compute_eval benchmark and resource server |
@ananthsub, @bxyu-nvidia |
review requested |
2026-06-02 00:07 UTC |
2026-06-03 00:07 UTC |
74d 13h |
| #1560 — docs: clarify resources server example-data contract and license validation |
@cwing-nvidia, @lbliii |
review requested |
2026-06-10 12:04 UTC |
2026-06-11 12:04 UTC |
68d 1h |
| #1541 — Add SC_Bench benchmark |
@adil-a, @ananthsub, @bxyu-nvidia, @cwing-nvidia |
review requested |
2026-06-13 00:57 UTC |
2026-06-15 00:57 UTC |
66d 12h |
| #1598 — Eshachar/openclaw swe |
@sdevare-nv |
review requested |
2026-06-16 18:19 UTC |
2026-06-17 18:19 UTC |
63d 19h |
| #1653 — docs: remove model recipe column from RL framework compatibility table |
@cwing-nvidia |
review requested |
2026-06-23 03:23 UTC |
2026-06-24 03:23 UTC |
59d 10h |
| #1752 — fix: handle vllm context length errors in nano v3 recipe |
@yfw |
review requested |
2026-06-26 05:53 UTC |
2026-06-29 05:53 UTC |
56d 7h |
| #2026 — feat: Switchyard token-capture integration for OpenHands SWE runs (zero-fork) |
@bxyu-nvidia |
review requested |
2026-07-14 22:36 UTC |
2026-07-15 22:36 UTC |
43d 14h |
| #1964 — feat(model-server): add OpenAI-compatible /v1/embeddings endpoint |
@ffrujeri |
review requested |
2026-07-15 16:18 UTC |
2026-07-16 16:18 UTC |
42d 21h |
| #1727 — fix(vllm_model): handle explicit null metadata in preprocessing |
@ananthsub |
review requested |
2026-07-15 17:11 UTC |
2026-07-16 17:11 UTC |
42d 20h |
| #2048 — feat: sandbox_agent swe resources server |
@adil-a, @ananthsub, @bxyu-nvidia, @hemildesai |
review requested |
2026-07-15 21:13 UTC |
2026-07-16 21:13 UTC |
42d 16h |
| #2052 — hermes_agent: repin to upstream hermes and fix the wrapper for its API |
@Glorf |
review requested |
2026-07-17 22:38 UTC |
2026-07-20 22:38 UTC |
40d 14h |
| #2155 — [GDPVal] Fix Stirrup reference paths and make head-server wait configurable |
@agronskiy |
review requested |
2026-07-29 16:40 UTC |
2026-07-30 16:40 UTC |
32d 20h |
| #2160 — [GDPVal] Fix what reaches the rubric judge, and stop silent zeros |
@agronskiy |
review requested |
2026-07-29 16:40 UTC |
2026-07-30 16:40 UTC |
32d 20h |
| #2162 — [GDPVal] Add model-agnostic Slurm orchestration for multi-allocation rollouts |
@agronskiy |
review requested |
2026-07-29 16:40 UTC |
2026-07-30 16:40 UTC |
32d 20h |
| #2304 — default port range to the 5000-5999 band |
@adil-a |
review requested |
2026-08-04 13:48 UTC |
2026-08-05 13:48 UTC |
28d 23h |
| #2311 — docs: rewrite Configure Agents index and document external agent harnesses |
@cwing-nvidia, @lbliii |
review requested |
2026-08-04 14:30 UTC |
2026-08-05 14:30 UTC |
28d 22h |
| #2386 — feat: harden SWE against filesystem overflow |
@sdevare-nv |
review requested |
2026-08-06 17:31 UTC |
2026-08-07 17:31 UTC |
26d 19h |
| #2142 — Add EnterpriseOps-Gym benchmark: resources server, benchmark registration, and per-turn telemetry agent |
@bxyu-nvidia, @cwing-nvidia |
review requested |
2026-08-08 08:52 UTC |
2026-08-10 08:52 UTC |
26d 4h |
| #2481 — feat: swe_trace_converter change for the opencode backend |
@bxyu-nvidia, @sdevare-nv |
review requested |
2026-08-11 18:44 UTC |
2026-08-12 18:44 UTC |
23d 18h |
| #2458 — feat: deepswe environment |
@Glorf, @hemildesai |
review requested |
2026-08-12 04:41 UTC |
2026-08-13 04:41 UTC |
23d 8h |
| #2346 — feat(sandbox): add Tenki provider |
@hemildesai |
review requested |
2026-08-12 16:45 UTC |
2026-08-13 16:45 UTC |
22d 20h |
| #2592 — Fix compute_pass_majority_metrics: score unanswered rollouts as failures in pass@k |
@gwarmstrong, @jkyi-nvidia |
review requested |
2026-08-18 16:55 UTC |
2026-08-19 16:55 UTC |
18d 20h |
| #2703 — docs(verifier): require verify to return the scored response |
@mpatel31415 |
review requested |
2026-08-24 05:09 UTC |
2026-08-25 05:09 UTC |
15d 8h |
| #2673 — fix: preserve structured MCP tool identity for verification |
@adil-a, @bxyu-nvidia |
review requested |
2026-08-25 00:10 UTC |
2026-08-26 00:10 UTC |
14d 13h |
| #2742 — chore(hermes_agent): use pinned upstream Hermes release |
@Glorf, @ananthsub, @laszkiewiczp |
review requested |
2026-08-25 03:18 UTC |
2026-08-26 03:18 UTC |
14d 10h |
| #2733 — feat(openclaw_agent): consume rollout-scoped Gym MCP metadata |
@cmunley1, @ffrujeri |
review requested |
2026-08-25 04:28 UTC |
2026-08-26 04:28 UTC |
14d 8h |
| #2744 — feat(nooa-agent): add safe declarative configuration |
@ffrujeri |
review requested |
2026-08-25 06:26 UTC |
2026-08-26 06:26 UTC |
14d 7h |
| #2745 — feat(nooa-agent): embed Gym-backed runtime |
@ffrujeri, @furgalep |
review requested |
2026-08-25 06:26 UTC |
2026-08-26 06:26 UTC |
14d 7h |
| #2746 — feat(nooa-agent): project resource evidence |
@ffrujeri |
review requested |
2026-08-25 06:26 UTC |
2026-08-26 06:26 UTC |
14d 7h |
| #2749 — docs(nooa-agent): add runnable workflows |
@ffrujeri |
review requested |
2026-08-25 06:26 UTC |
2026-08-26 06:26 UTC |
14d 7h |
| #2771 — feat: add raw_rewards_and_lengths to group-level reward profile metrics |
@bxyu-nvidia, @jkyi-nvidia |
review requested |
2026-08-25 21:53 UTC |
2026-08-26 21:53 UTC |
13d 15h |
| #2783 — refactor: decouple rollout identity from capture |
@Glorf, @ffrujeri, @pthombre |
review requested |
2026-08-26 08:14 UTC |
2026-08-27 08:14 UTC |
13d 5h |
| #2787 — fix(token-id-capture): classify auxiliary model requests |
@cmunley1 |
review requested |
2026-08-26 11:20 UTC |
2026-08-27 11:20 UTC |
13d 2h |
| #2758 — feat: sandbox provider image prepare hook + update prepare step in AnyTerminal |
@hemildesai |
review requested |
2026-08-26 16:17 UTC |
2026-08-27 16:17 UTC |
12d 21h |
| #2732 — chore(data): apply the dataset decoupling migration (agent_ref out of data files) |
@ananthsub |
review requested |
2026-08-26 23:21 UTC |
2026-08-27 23:21 UTC |
12d 14h |
| #2866 — Add support for judging and rejudging saved responses |
@bxyu-nvidia |
review requested |
2026-08-29 00:07 UTC |
2026-08-31 00:07 UTC |
11d 13h |
| #2879 — feat(sandbox): IdeGYM sandbox provider |
@hemildesai |
review requested |
2026-08-31 23:09 UTC |
2026-09-01 23:09 UTC |
9d 14h |
| #2904 — fix(cli): report unresolved config interpolations cleanly in manifest validation |
@Glorf |
review requested |
2026-09-01 05:34 UTC |
2026-09-02 05:34 UTC |
9d 7h |
| #2870 — fix(agent): unify context compaction trace contract |
@ananthsub |
review requested |
2026-09-02 20:27 UTC |
2026-09-03 20:27 UTC |
7d 17h |
| #3050 — fix(token-id-capture): fingerprint the served response dialect |
@pthombre |
review requested |
2026-09-03 08:30 UTC |
2026-09-04 08:30 UTC |
7d 4h |
| #3058 — fix(code-gen): replace Manager proxies with bounded IPC |
@fgalko-oss |
review requested |
2026-09-03 19:21 UTC |
2026-09-04 19:21 UTC |
6d 18h |
| #3078 — docs: update NeMo RL tutorial container |
@cwing-nvidia, @kajalj22, @lbliii, @sephmard |
review requested |
2026-09-04 00:19 UTC |
2026-09-07 00:19 UTC |
6d 13h |
| #3047 — fix(model-server): preserve aliases in ORJSON responses |
@kajalj22, @pthombre |
review requested |
2026-09-04 18:28 UTC |
2026-09-07 18:28 UTC |
5d 19h |
| #2015 — harness agent |
@ananthsub |
review requested |
2026-09-04 21:07 UTC |
2026-09-07 21:07 UTC |
5d 16h |
| #3117 — feat(server_utils): opt-in CRLF heartbeat so long HTTP/1.1 model calls survive Global Accelerator's 340s idle timeout |
@ananthsub, @kajalj22 |
review requested |
2026-09-08 04:58 UTC |
2026-09-09 04:58 UTC |
4d 8h |
| #3133 — docs: standardize verification task data convention |
@adil-a |
review requested |
2026-09-09 12:16 UTC |
2026-09-10 12:16 UTC |
3d 1h |
| #3189 — fix(responses): preserve refusal conversion semantics |
@bxyu-nvidia, @cmunley1 |
review requested |
2026-09-09 14:04 UTC |
2026-09-10 14:04 UTC |
2d 23h |
| #2954 — docs: document gym env prefetch CLI #2937 |
@ritaneves |
review requested |
2026-09-09 16:54 UTC |
2026-09-10 16:54 UTC |
2d 20h |
| #2580 — docs: fix Fern polish issues from NVBugs 6570461 |
@cwing-nvidia, @sephmard |
review requested |
2026-09-09 16:55 UTC |
2026-09-10 16:55 UTC |
2d 20h |
| #3306 — fix(server-utils): disable uvicorn proxy headers by default |
@ananthsub |
review requested |
2026-09-12 01:00 UTC |
2026-09-14 01:00 UTC |
1d 12h |
Pull request handoff SLA tracker
Review during the daily stand-up alongside the architecture/RFC decision tracker.
The SLA is one business day per handoff; weekends are excluded and public holidays are not modeled.
Drafts, bots, and PRs labeled
stale,paused, ordo-not-trackare excluded.Last refreshed: 2026-09-15T13:28+00:00
Newly breached (7)
Waiting on reviewers (4)
Needs review assignment (3)
Older breach backlog (121)
Waiting on reviewers (57)
Waiting on authors (5)
Needs review assignment (59)