ginlix-ai — LangAlpha
LangAlpha through 2026.09.07 contains an indirect prompt injection vulnerability that allows remote attackers to execute shell commands or Python by planting instructions in web content retrieved by the agent. Because Bash and ExecuteCode lack an allow-list and human-in-the-loop approval, attackers can run code in the workspace sandbox and exfiltrate vault secrets from _internal/.vault_secrets.json.
Available for technologies you monitor inside Seraph CTI.
Surfaced automatically when a documented actor is linked to this CVE.
Shown when this vendor/product is relevant to your industry.
Campaign and indicator context, where Seraph analysts have published it.
Declare your technology stack inside Seraph CTI and get notified the moment a vulnerability matches it.
