Skip to content

Tags: benchflow-ai/skillsbench

Tags

v1.1

Toggle v1.1's commit message
SkillsBench v1.1 — 87-task native task.md release (re-stamped)

Snapshot commit: b63b7b2
- 12 broken oracles fixed self-contained (post-#720 skills-not-injected)
- exoplanet-detection-period: TLS use_threads pinned (no OOM on many-core Docker)
- fix-erlang-ssh-cve: build parallelism capped
- All 87 v1.1 registry entries re-pinned to merge commit 55bfe69 with recomputed digests
- BenchFlow aligned to >=0.6.3,<0.7

v1.0

Toggle v1.0's commit message
SkillsBench v1.0 — 87-task benchmark snapshot