Description
PraisonAI before 4.6.78 contains a prompt injection defense bypass vulnerability where the injection defense only blocks threats classified as CRITICAL, requiring three or more detector families to match simultaneously. Attackers can craft single or double-vector prompt injections that are classified as HIGH threat level and pass through unblocked to reach the model.
Affected products
| Vendor | Product | Affected | Fixed |
|---|---|---|---|
| MervinPraison | PraisonAI | < 4.6.78 | 4.6.78 |
Affected packages
| Ecosystem | Package | Affected | Fixed | Record |
|---|---|---|---|---|
| PyPI | praisonai | < 4.6.78 | 4.6.78 | GHSA-4r3p-w3mc-5v34 |
References
- GitHub Security Advisory (GHSA-4r3p-w3mc-5v34) vendor-advisory
- VulnCheck Advisory: PraisonAI before 4.6.78 Prompt Injection Defense Bypass third-party-advisory