CVE-2026-78379Patch

LOWCVSS 9.2 · CRITICAL

Signal is active with 1 mentions in latest observed window

Immediate actions

Patch affected systems immediately

Recommended action window: Monitor and triage in normal cycle

NVD description

Improper neutralization of input used for LLM prompting in the python_repl tool in Amazon Strands Agents Tools before 0.8.5 might allow remote actors to execute arbitrary Python code on the agent's host by bypassing the human consent gate, via a crafted prompt that forwards non_interactive_mode as a keyword argument through the batch tool. To remediate this issue, users should upgrade to version 0.8.5 or later.

Back to dashboard

0.5/ 10 priority

Sources & remediation

Other references

Priority

LOW

Exploitation

NONE

PoC

NONE

Patch

AVAILABLE

Momentum

NONE

Threat summary

Patch or workaround signal is available
1 mentions across 1 observed day

What's happening

Patch or workaround mentioned in 1 signal
Technical details provided in 1 signal
1 total mentions across 1 day

Deep dive

Activity timeline1 mentions / 1d

Signal classification1 categories

Patch

1100.0%

Referenced assets1 URL

By indicator

https://aws.amazon.com/security/security-bulletins/2026-089-aws/

Full discourse1 post

Swif@swif_ai46d
Patch
AWS fixed CVE-2026-78379: chaining tools let a crafted prompt make python_repl skip its human approval gate and run code on the agent's host. Approval flags must come from the operator only, never the model. https://aws.amazon.com/security/security-bulletins/2026-089-aws/
Post summary
AWS released a patch for CVE-2026-78379; the vulnerability involves crafted prompts that bypass python_repl’s approval gate, allowing code execution on the agent host, and detailed information is available in the AWS security bulletin.
0000041
59 followersView on X

Explore more

Back to dashboard