HighVulnerabilityLLM-specific
GHSA-8hp3-xxrg-v9c8: CodeWhale: Automation auto_approve defaults to true, bypassing the task-path hardening fix for silent recurring agent execution
- Identifiers
- CVE-2026-83817GHSA-8hp3-xxrg-v9c8
- Published
- Record updated
- Affected
- codewhale <= 0.8.67
- codewhale-tui <= 0.8.67
- Fixed in
- 0.9.0
- codewhale 0.9.0 released July 16, 2026, 126 days before disclosure
- codewhale-tui 0.9.0 released July 16, 2026, 126 days before disclosure
- Known exploitation
- Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
- EPSS
- 0.7%
Summary
CodeWhale 0.8.67 and earlier are affected by GHSA-8hp3-xxrg-v9c8. The automation subsystem resolves an omitted auto_approve to true, so a single approved automation_create yields a durable, recurring agent that auto-approves every tool call without a human present, bypassing the task-path hardening from CVE-2026-45374.
Mitigation
Fixed in 0.9.0 (commit 208421b9d23d81cea3311a1eeb2a2ea81d637686). Users should upgrade to 0.9.0 or later.
Related items
- HighGHSA-3mgh-93r5-r7jf: CodeWhale: exec_shell parallel-readonly classifier auto-approves fd/rg/git grep exec flags enabling silent RCESimilar attack · GitHub Advisory Database
- MediumCVE-2026-18883: ServiceNow Email Response AI Agent indirect prompt injection flawSimilar attack · NVD/CVE Database
- InfoAI-Defined Vehicles: Securing the Decisions That Drive Vehicle BehaviorSimilar attack · VicOne Blog
- HighCVE-2026-75612: yt-mcp arbitrary file read through add_attachment file_pathSimilar attack · NVD/CVE Database
- InfoPIPETTE: Prompt Injection Purple-Teaming for Robust LLM AgentsSimilar attack · Arxiv (cs.CR + cs.CL + cs.LG)