MediumVulnerabilityLLM-specific

CVE-2026-108592: mini-swe-agent information exposure in BubblewrapEnvironment sandbox

Source: NVD(opens in a new tab)
Identifier: CVE-2026-108592
Published: October 10, 2026
Record updated: October 10, 2026

Known exploitation: Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.

Summary

mini-swe-agent versions 1.10.0 through 2.4.6 contain an information exposure flaw in BubblewrapEnvironment. The bwrap invocation omits --clearenv, so sandboxed commands inherit the host environment. An attacker using prompt injection in processed task content can make the agent read API keys from that environment and exfiltrate them over the shared network.

Mitigation

The source does not state a fix yet. Check the original advisory for updates.

Details

Scores and exploitation

CVSS score: 5.3(medium)
CVSS vector: CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:N/A:N
Attack vector: network
Attack complexity: high
Privileges required: none
User interaction: required
EPSS: Not scored. FIRST had not published an EPSS score for this CVE when it was last checked on October 10, 2026.
Disclosure date: October 10, 2026

Weaknesses and techniques

Weakness (CWE): CWE-526
AI threat technique (MITRE ATLAS): AML.T0051

Classification

Attack type: Prompt injection,
Data extraction
Attack sophistication: Moderate
AI component targeted: Agent
Impact: Confidentiality
Labels: security

How this record was produced

Summaries, mitigations and labels on this page are generated automatically from the source. The linked source is authoritative. The source gives this advisory a description and no title, so the title above is a headline generated from the source text.

Original source: https://nvd.nist.gov/vuln/detail/CVE-2026-108592

First tracked: October 10, 2026 at 08:09 PM ET, from NVD/CVE Database

Classifier: claude-haiku-5-5, prompt v4

Classifier confidence: 95% (self-assessed)

Summary

Mitigation

Related items