exploit-verifier
Warn
Audited by Socket on Jun 23, 2026
2 alerts found:
SecurityAnomalySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
Anomalyreferences/ai-payloads/llm-injection-payloads.md
LOWAnomalyLOW
references/ai-payloads/llm-injection-payloads.md
This file is not executable software; it is a static, highly actionable reference of LLM prompt-injection/evasion/exfiltration payloads and verification steps. While no direct malware behavior is present in the fragment, its inclusion as a distributed dependency is security-sensitive because it materially enables misuse against LLM-based systems with guardrails or tools. Treat as a high-content-misuse risk artifact; further review should confirm whether any importing code exists elsewhere in the package and how (or whether) these strings are used at runtime.
Confidence: 74%Severity: 62%
Audit Metadata