AI Stack Alert High CVE-2026-115415 For AI Eng, AppSec, SOC / IR, CISO AI-assisted
CVE-2026-115415: Strands Agents Tools environment tool prompt injection exposes host secrets
Strands Agents Tools versions before 0.8.6 have an improper neutralization of input used in LLM prompting in the environment tool. A remote actor can use a crafted prompt to disable the tool's secret masking and read the agent host's environment variables in cleartext, including cloud credentials and API keys.
Why it matters
Any agent host running the vulnerable environment tool can have its cloud credentials and API keys exposed to a remote actor who can supply prompts to the agent. This can lead to unauthorized access to connected cloud and third-party services.
Do this
- Upgrade Strands Agents Tools to version 0.8.6 or later on all agent hosts.
- Rotate all credentials, cloud keys, and API keys present in the environment of any agent that loaded the environment tool before the upgrade.
- Identify agents that load the environment tool and review whether they accept untrusted prompt input.
- Review agent logs for prompts that reference environment variables or attempt to disable secret masking, and review cloud audit logs for use of credentials from affected hosts.
Exploit Clock
Exploitation unlikely (for now)
- EPSS 0.34%
Signals
- NoCISA KEV
- 5.9CVSS
- 0.34%EPSS
- NoPublic PoC
- 34/100Priority
- 70%Confidence
Evidence
“Improper neutralization of input used for LLM prompting in the environment tool in Strands Agents Tools before 0.8.6 might allow remote actors to obtain the agent host's environment variables in cleartext”
“users should upgrade to version 0.8.6 or later and rotate any credentials that were present in the environment of agents that loaded the environment tool.”
Sources
Affected / related
Strands Agents Tools, Prompt injection
Published 2026-10-28 20:17 UTC · updated 2026-11-08 02:37 UTC · AI-assisted summary, checked against quoted sources · How we score