Back to skill
Skillv1.0.32
Static analysis security
Agentshield Audit · Deterministic local checks for risky code patterns and metadata mismatches.
Scanner verdict
SuspiciousApr 1, 2026, 1:37 PM
- Summary
- Detected: suspicious.prompt_injection_instructions
- Reason codes
- suspicious.prompt_injection_instructions
- Engine
- v2.2.0
Evidence
warnSKILL.md:366
Prompt-injection style instruction pattern detected.
suspicious.prompt_injection_instructions