Other documented method
Last commit on April 29, 2026·Created on April 7, 2026
Fangcun-AI/SkillWard
“Audits third-party skill bundles before installation.”
Combined rank
#513
across all skills
In Security
#25
category rank
Stars
130
+0.8% in last 7d
Forks
5
0.0% in last 7d
Watchers
3
+50.0% in last 7d
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of real attention, adoption, and active maintenance.
Description
SkillWard Audit is a thin client for the SkillWard public scanning service. Given a local skill bundle — a folder with SKILL.md, or a .zip or .tar.gz archive — it packages the input, uploads it to the remote service, streams scan progress, and returns a verdict: SAFE, MEDIUM RISK, or HIGH RISK. The report includes warnings with critical and warning levels, surfaced from static analysis, LLM review, and an optional Docker sandbox. It helps AI agent power users, platform teams, and security engineers vet external skills before installing them, and provides a documented safety check that can be archived for compliance.
WHO IT'S FOR
AI Agent Power Users
vet third-party skills before installation
Platform / DevEx Teams
establish a security gate for skill ingestion
Security Engineers (AI/ML)
audit skill bundles for hidden threats
AI Governance / Compliance Teams
review third-party skill safety reports
30 Days of GitHub Stars
+3 stars↑ 0.8% monthly change
Jul 7Jul 14Jul 21Jul 28Aug 5
Compatibility & Install1 Skill · 2 Compatible agents · 2 categories · 1 install path
Compatible agents
❯Claude Code❯OpenClaw
Categories
SecurityDeveloper Tooling
Install via
Repository instructions
Also in this category4 of 24