If you find vskill useful, give it a star on GitHub→
vskill
LoginSkillsStudioWatchSubmitPublishersTrustQueueDocsInsights

Documentation

Everything you need to install, verify, and publish AI agent skills.

Get Started

Quick Startv1.0.18

Install vskill, run your first security scan, and understand the three-tier trust model.

Skill StudioLocal IDE

Local IDE for skills — author, A/B eval, sweep across models, publish in one click. Open with `npx vskill@latest studio`.

CLI Reference

Complete command reference with flags, examples, and usage patterns for every vskill command.

Security

Security Guidelinesv1.0

52 scan patterns, 9 threat categories, scoring algorithm, and a pre-submission checklist.

Submit a Skill

How to write SKILL.md, pass verification, and publish to the registry.

Ecosystem

Plugin Marketplace16 skills

16 expert skills across 9 domain plugins — frontend, marketing, mobile, productivity, and more.

Trust Center

Browse verified, certified, and blocked skills. View the full blocklist and report suspicious skills.

FAQ

Common questions about trust tiers, false positives, DCI blocks, and troubleshooting.

SpecWeave Integrationv1.0.585

vskill is the package manager for SpecWeave. Install, update, and audit skills in your projects.

Quick Install

bash
npx vskill@latest install remotion-best-practices

Quick Verify

bash
npx vskill@latest scan .

Run from any directory containing a SKILL.md to get a tier-1 security score (52 pattern checks).

Insights

Claude Code Source ExposedAnalysis

512K lines analyzed — architecture deep dive, hidden features, and security audit.

7 Hidden Limits & CLAUDE.md OverridesPractical

File read caps, tool truncation, context spirals — every claim verified against the source.

54
Agent platforms
52
Scan patterns
16
Expert skills
3-tier
Verification
GitHub→npm→SpecWeave→
vskill

Securing the AI skills supply chain.

SkillsStudioWatchPublishersTrustDocsQueueSubmitInsightsGitHubnpmSpecWeave

MIT License