skillsbenchbenchflow-ai/skillsbench
[Core Features]: The first benchmark for how effectively AI agents use skills, evaluating both skill effectiveness (qual…
InstallOne-line install
UnconfirmedWhether this repository follows the Agent Skills spec (that is, whether it contains a SKILL.md) has not been verified yet. Until that is settled, clone it and use what you need:
git clone https://github.com/benchflow-ai/skillsbench.git
OverviewOverview
[Core Features]: The first benchmark for how effectively AI agents use skills, evaluating both skill effectiveness (quality of instructions, scripts and resources in a skill folder) and agent behavior through gym-style benchmarking. It aims to build the broadest, highest-quality benchmark with tasks requiring two or more composed skills where SOTA stays below 50%, targeting GPT, Claude, Gemini, GLM, Kimi and MiniMax. It ships a BenchFlow CLI, a Hugging Face dataset and a website. [Applicable Scenarios]: Agent skill effectiveness evaluation, model comparison, skill research and benchmark contribution. [Target Roles]: AI researcher, model evaluation engineer, agent developer.
TopicsTopics
RatingUser ratings
Nobody has rated this skill yet. Yours would be the first on the site — your vote becomes the initial average.
RepositoryRepository
| Skill ID | 488 |
|---|---|
| Data source | GitHub |
| Repository | benchflow-ai/skillsbench |
| Category | AI & Agent Skills / Agent Skill Evaluation Benchmark |
| Primary language | PDDL |
| License | Apache License 2.0 |
| Default branch | main |
| GitHub repository ID | 1124479538 |
| Node ID | R_kgDOQwYyMg |
| Rank | #531 (position within the source ranking) |
| Site-wide popularity rank | #614 (computed live from Stars, across all 5875 entries) |
GrowthStar / Fork growth
One snapshot per day, tracking how Stars and Forks changed
- Now1,927
- Previous point · 10-11 1,837
- vs previous point +90
- 3 data points 51 days
- Now378
- Previous point · 10-11 371
- vs previous point +7
- 3 data points 51 days
TimelineTimeline
CommitLatest commit
Committed by bingran-you. Update Category: Refactoring & Structural Improvements. Commit details: Rename task metadata sections to sandbox (#1037). This update restructures internal code for better maintainability, without behavior changes.