Local file disclosure via prompt injection
Published Aug 4, 2024 · Updated Aug 30, 2024
Prompt injection in Stition Devika at commit cdfb782b0e634b773b10963c8034dc9207ba1f9f allows remote attackers to read local files. The Gemini 1.0 Pro integration sets BLOCK_NONE for hate-speech and harassment categories in safety_settings, disabling those content filters before prompt-derived commands are processed. The network-accessible chat supplies the malicious prompt, and successful exploitation discloses files readable by the Devika process.
Summary
What happened
Prompt injection in Stition Devika at commit cdfb782b0e634b773b10963c8034dc9207ba1f9f allows remote attackers to read local files. The Gemini 1.0 Pro integration sets BLOCK_NONE for hate-speech and harassment categories in safety_settings, disabling those content filters before prompt-derived commands are processed. The network-accessible chat supplies the malicious prompt, and successful exploitation discloses files readable by the Devika process.
The record
- CVE
- CVE-2024-6331
- Published
- Aug 4, 2024
- Updated
- Aug 30, 2024
- Vendor
- Stition
- Product
- Devika
- Classifications
- CWE-74, T1005
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- Aug 4, 2024CVE publishedPublication date reported by the CVE source.
- Aug 30, 2024Record 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 <=latest
- Affected versionversion=unspecified <=latest
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
- Huntr prompt-injection file-read 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