Skip to content

Benchmark Results ​

Run npm run benchmark to reproduce on your machine and regenerate the SVG chart.

Environment: Node.js v24.18.0, macOS (darwin, x64) Fixture (local): SKILL.md + 1 JS file (2 files, 78 bytes) Fixture (GitHub): sametcelikbicak/coverage-guardIterations: 10 per tool per scenario Date: 2026-10-09

What these numbers do and do not measure ​

The ratio below is not a like-for-like comparison of install logic, and the gap is worth roughly two orders of magnitude:

  • npx skills is invoked as npx --yes skills add … on every iteration. That re-resolves and re-downloads the package from the registry each time, so its column includes npm package acquisition, not skill installation. With the package already cached locally it is substantially faster.
  • rolecraft is measured in-process. Its column calls installSkill() directly, so it excludes Node startup (~115 ms floor, ~290 ms as shipped) and process spawn.

Two runs on the same machine on the same day produced local ratios of 356x and 266x — a 34% spread from nothing but run-to-run variance and npm cache state. The local scenario is small enough that a few milliseconds of noise dominates the ratio. Read the absolute column values, and re-run before quoting a figure.

The install work itself is the part this project controls: single-digit to ~20 ms for a local skill. What npx skills spends that time on once acquisition is excluded is not measured here — adding that measurement would need a second run with the package pre-cached.

Benchmark comparison chart

Local path install ​

Toolavgminmaxp50vs rolecraft
rolecraft17.56 ms7.65 ms38.65 ms12.56 ms1.00x
skills (Vercel)4665.86 ms4288.63 ms4914.45 ms4677.49 ms265.74x
@agentskill.sh/cli————N/A

@agentskill.sh/cli is marketplace-only and does not support local paths.

GitHub install (sametcelikbicak/coverage-guard) ​

Toolavgminmaxp50vs rolecraft
rolecraft2553.22 ms2207.98 ms3065.89 ms2587.09 ms1.00x
skills (Vercel)8219.69 ms6981.81 ms8909.28 ms8429.06 ms3.22x
@agentskill.sh/cli5432.83 ms4594.18 ms6822.82 ms5501.85 ms2.13x

@agentskill.sh/cli completed this run. Earlier runs of this benchmark recorded it as failing during agent detection, so treat its row as version-dependent rather than settled.

Key takeaways ​

ScenariorolecraftVercel skills@agentskill.sh/cli
Local skill install✅ 17.6 ms✅ 4666 ms❌ not supported
GitHub skill install✅ 2.6 s✅ 8.2 s✅ 5.4 s
Zero dependencies✅ 0❌ 1 dep❌ 2 deps