CVE-2026-57120 PraisonAI: execute_code sandbox bypass: str.format C-level attribute access reads every blocklisted dunder
Attackers can bypass sandbox restrictions in PraisonAI to read sensitive internal attributes via string formatting when auto-approval is enabled.
- Exploited in the wild
- Not known
- Login needed
- Any signed-in user
- Attackable
- Over the internet
- Victim action needed
- Not needed
- Fix
- Available
What is CVE-2026-57120?
CVE-2026-57120 is a medium-severity code injection vulnerability in MervinPraison praisonaiagents, published on 14 Sep 2026. Any signed-in user can exploit it over the network, with no action from the victim.
Official description
PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.59, execute_code sandbox mode permits runtime assembly of blocklisted dunder names and allows str.format or str.format_map to resolve dotted fields through C-level attribute access that bypasses _safe_getattr. This exposes class, qualified-name, base-class, globals, and object-dictionary attributes to prompt-influenced code when approval is automatically granted, producing a high-impact read primitive without establishing a complete in-process execution chain. This issue is fixed in praisonaiagents 1.6.59.
Weakness: CWE-693
Which versions are affected by CVE-2026-57120?
3 open source packages ship an affected version, including praisonaiagents.
| Package | Affected | Fixed in |
|---|---|---|
praisonaiagents PyPI | All versions | 1.6.59 |
praisonaiagents PyPI | All versions | 1.6.59 |
praisonaiagents PyPI | All versions | 1.6.59 |
How do I fix CVE-2026-57120?
Upgrade praisonaiagents to 1.6.59 or later.
Upgrade, and never evaluate templates, expressions or code built from user input. More about Code injection.
Is CVE-2026-57120 being exploited?
It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-57120 exists.
Sources
Also known as GHSA-pv2j-rghr-v5r9, PYSEC-2026-3533.