Skip to content

CVE-2026-45311 CodeWhale: run_tests Tool Enables RCE via Malicious Repository Without Approval

CriticalCVSS 9.6 EPSS 0.57%
Published
Updated
Exploit status
No public exploit known
CISA SSVC
Proof of concept public · automatable: no · technical impact: total

Description

CodeWhale is a DeepSeek + MiMo coding agent in terminal. From 0.3.0 to 0.8.23, the run_tests tool executes cargo test in the workspace with ApprovalRequirement::Auto, meaning it runs without any user approval prompt. cargo test compiles and executes arbitrary code: test binaries, build.rs build scripts, and proc macros. While auto-approving test execution is a deliberate design choice, it creates an inconsistency in the security boundary. However, in a malicious repository, test code can execute arbitrary shell commands, exfiltrate credentials, or establish persistence with zero approval. The attack is amplified by AGENTS.md (auto-loaded into the system prompt), which can instruct the model to run tests proactively at session start. This vulnerability is fixed in 0.8.23.

Upgrade path

4 packages: 3 fixed.

PackageAffectedFixed inStatus
crates.io deepseek-tui(crates.io) >= 0.3.0, < 0.8.23 0.8.23 Fixed
crates.io deepseek-tui-cli(crates.io) >= 0.3.0, < 0.8.23 0.8.23 Fixed
npm deepseek-tui(npm) >= 0.3.0, < 0.8.23 0.8.23 Fixed
Hmbown CodeWhale = >= 0.3.0, < 0.8.23 No fix listed Versions listed

Affected products

ProductAffected versionsFixed in
Hmbown CodeWhale = >= 0.3.0, < 0.8.23 No fix listed
crates.io deepseek-tui(crates.io) >= 0.3.0, < 0.8.23 0.8.23
crates.io deepseek-tui-cli(crates.io) >= 0.3.0, < 0.8.23 0.8.23
npm deepseek-tui(npm) >= 0.3.0, < 0.8.23 0.8.23

References