# Skill Retrieval on Terminal-Bench

Source: [Skill Retrieval on Terminal-Bench](<https://www.sota2.com/research/sota/skill-retrieval-on-terminal-bench>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-31T06:59:53.536Z

Evaluation of the model's ability to retrieve the correct skill using the Terminal-Bench dataset.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Skill Retrieval |
| Dataset (including any stated split) | Terminal-Bench |
| Metric | Mean Skills Retrieved per Task |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Mean Skills Retrieved per Task](<https://www.sota2.com/research/sota/skill-retrieval-on-terminal-bench.md?metric=Mean%20Skills%20Retrieved%20per%20Task>): Higher is better; unit: SCALAR.

- [Tasks Retrieved Rate](<https://www.sota2.com/research/sota/skill-retrieval-on-terminal-bench.md?metric=Tasks%20Retrieved%20Rate>): Higher is better; unit: PERCENT.

- [Tasks Used Rate](<https://www.sota2.com/research/sota/skill-retrieval-on-terminal-bench.md?metric=Tasks%20Used%20Rate>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Mean Skills Retrieved per Task | Result date | Report paper |
| --- | --- | --- | --- |
| SkillFlow | 1.5 | Not recorded | 68fa77dc1e209131cea759e1 |
| Vercel | 0.9 | Not recorded | 68fa77dc1e209131cea759e1 |
| SkillFlow | 0.9 | Not recorded | 68fa77dc1e209131cea759e1 |

### SkillFlow

Mean Skills Retrieved per Task: 1.5. Result ID: 69cb6f5747f34ce924b33297.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77dc1e209131cea759e1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77dc1e209131cea759e1.md>)

Recorded evaluation configuration: {"Condition":"full"}. This record does not assert independent reproduction.

### Vercel

Mean Skills Retrieved per Task: 0.9. Result ID: 69cb6f5747f34ce924b33295.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77dc1e209131cea759e1.md>)

Recorded evaluation configuration: {"Condition":"Vercel"}. This record does not assert independent reproduction.

### SkillFlow

Mean Skills Retrieved per Task: 0.9. Result ID: 69cb6f5747f34ce924b33296.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77dc1e209131cea759e1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77dc1e209131cea759e1.md>)

Recorded evaluation configuration: {"Condition":"top-1"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.