skill-test
Fail
Audited by Socket on Mar 10, 2026
1 alert found:
Obfuscated FileObfuscated FileSKILL.md
HIGHObfuscated FileHIGH
SKILL.md
Overall, the skill-test framework appears benign and proportionate to its stated purpose of automating discovery, testing, and reporting for Claude Code skills. The main concerns relate to potential autonomous execution of external skills (autonomy_abuse) and a minimal supply-chain risk if transitive installs are performed without verification. These should be mitigated by sandboxing, explicit user confirmations for installations, and auditing of any externally fetched skills. Absent explicit credential handling or network exfiltration, the footprint remains consistent with a developer tooling/quality assurance utility.
Confidence: 98%
Audit Metadata