SkillBench
PassAudited by ClawScan on May 1, 2026.
Overview
The skill is coherent for benchmarking other skills, but users should notice that it installs an external CLI, can sync benchmark data to ClawVault, and documents recurring monitoring workflows.
This appears to be a purpose-aligned benchmarking skill. Before installing, verify the external npm package, avoid putting sensitive details in benchmark task names or error messages, and only enable ClawVault sync, monitoring, or cron-style workflows if you want persistent records and recurring activity.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing the skill means trusting the external npm package that supplies the skillbench command.
The skill depends on an external npm package that provides the CLI. This is expected for the stated purpose, but the package source is not included in the artifact set.
node | package: @versatly/skillbench | creates binaries: skillbench
Verify the npm package publisher and version before installing, especially because the registry source is listed as unknown.
Task names, timing, success/failure results, error types, and skill-performance history may be stored outside the local session and used later.
Benchmark records may be persisted or reused through ClawVault, which the artifact describes as a memory system for AI agents.
Benchmarks sync to [ClawVault](https://clawvault.dev) automatically.
Avoid recording sensitive task names or private operational details unless you are comfortable syncing them to ClawVault; review ClawVault storage and retention settings.
If enabled, recurring monitoring could keep running tests and recording results over time.
The skill documents a recurring monitoring mode. It is presented as a user-run command, not as hidden persistence.
skillbench watch --interval 300 # Continuous monitoring every 5 min
Run monitoring or generated cron workflows only when you intend ongoing activity, and review any scheduled jobs before enabling them.
