Skip to content

CVE-2026-83816 CodeWhale: exec_shell parallel-readonly classifier auto-approves fd/rg/git grep exec flags enabling silent RCE

HighCVSS 8.1 EPSS 0.61%
Published
Updated
Exploit status
No public exploit known
CISA SSVC
No known exploitation · automatable: no · technical impact: total

Description

CodeWhale is a DeepSeek and MiMo coding agent for the terminal. Prior to 0.9.0, the ExecShellTool::approval_requirement_for path used is_parallel_readonly_command to downgrade commands to automatic approval based on a flag-stripped command allowlist. The allowlist treated fd, rg, and git grep as read-only even when fd -x or fd --exec, rg --pre, or git grep --open-files-in-pager selected an external program. registered_tool_approval_required returned without displaying a prompt for that automatic classification, and the later command-safety check also treated the base utilities as safe. Attacker-controlled content that influenced an agent tool call could therefore cause arbitrary program execution on the developer host without the approval promised by the default interactive profile, resulting in unauthorized access, modification, or disruption. This issue is fixed in version 0.9.0.

Upgrade path

3 packages: 2 fixed.

PackageAffectedFixed inStatus
crates.io codewhale-tui(crates.io) < 0.9.0; <= 0.8.67 0.9.0 Fixed
npm codewhale(npm) < 0.9.0; <= 0.8.67 0.9.0 Fixed
codewhale-hq CodeWhale = < 0.9.0 No fix listed Versions listed

Affected products

ProductAffected versionsFixed in
codewhale-hq CodeWhale = < 0.9.0 No fix listed
crates.io codewhale-tui(crates.io) <= 0.8.67 No fix listed
npm codewhale(npm) <= 0.8.67 No fix listed
crates.io codewhale-tui(crates.io) < 0.9.0 0.9.0
npm codewhale(npm) < 0.9.0 0.9.0

References