Code execution via auto-approved Python tool
Published Aug 18, 2026 · Updated Aug 18, 2026
Code injection in rlm_eval in Hmbown CodeWhale before 0.8.64 allows attackers to execute arbitrary Python on the host via prompt injection. The tool reads an LLM-controlled code field and passes it to a real python3 interpreter, while approval_requirement() overrides the default ExecutesCode safeguard with ApprovalRequirement::Auto so the engine skips approval checks and emits no approval prompt. Reachability requires the agent to process attacker-influenced content such as a web page, fetched URL, repository file, or MCP tool result, after which code runs with the local account's privileges and unsandboxed filesystem, process, and network access.
Summary
What happened
Code injection in rlm_eval in Hmbown CodeWhale before 0.8.64 allows attackers to execute arbitrary Python on the host via prompt injection. The tool reads an LLM-controlled code field and passes it to a real python3 interpreter, while approval_requirement() overrides the default ExecutesCode safeguard with ApprovalRequirement::Auto so the engine skips approval checks and emits no approval prompt. Reachability requires the agent to process attacker-influenced content such as a web page, fetched URL, repository file, or MCP tool result, after which code runs with the local account's privileges and unsandboxed filesystem, process, and network access.
The record
- CVE
- CVE-2026-75858
- Published
- Aug 18, 2026
- Updated
- Aug 18, 2026
- Vendor
- Shannon Labs
- Product
- CodeWhale
- Classifications
- CWE-94, T1059.006
- Attack vector
- local
- Privileges
- unauthenticated
Timeline
How it unfolded
- Aug 18, 2026CVE publishedPublication date reported by the CVE source.
- Aug 18, 2026Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=0.8.33 <0.8.41
- Affected versionversion=0.8.41 <0.8.64
What conditions does exploitation require?
What is affected?
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- rlm_eval approval-policy bypass proof of conceptproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo