CVE-2026-47395: CWE-200: Exposure of Sensitive Information to an Unauthorized Actor in MervinPraison PraisonAI
PraisonAI is a multi-agent teams system. Prior to version 4.6.40 of PraisonAI, corresponding to version 1.6.40 of praisonaiagents, PraisonAI's direct-prompt CLI automatically expands `@url:` mentions in raw prompt text before agent execution begins. If a prompt contains `@url:<http-or-https-url>`, the CLI calls `MentionsParser.process(...)`. The `@url:` handler then performs a direct `urllib.request.urlopen()` request to the attacker-controlled URL and returns the response body. That response body is prepended to the final model prompt context. There is no loopback/private-address restriction, no metadata-service restriction, and no approval gate before the fetch. As a result, attacker-influenced prompt text can cause the operator's machine to fetch localhost-only HTTP resources and inject the response into model context. PraisonAI version 4.6.40 and praisonaiagents version 1.6.40 contain a fix.
AI Analysis
Technical Summary
PraisonAI's direct-prompt CLI feature automatically processes '@url:' mentions in raw prompt text by performing an HTTP(S) request to the specified URL using urllib.request.urlopen(). There are no restrictions on the requested URL, allowing attacker-controlled prompts to cause the operator's machine to fetch localhost or other internal network resources. The fetched response body is then prepended to the model prompt context, exposing sensitive internal information to unauthorized actors. This vulnerability is identified as CWE-200 (Exposure of Sensitive Information). Versions prior to 4.6.40 of PraisonAI and 1.6.40 of praisonaiagents are affected. The issue is fixed starting from these versions.
Potential Impact
An attacker who can influence prompt text can cause the operator's machine to make HTTP requests to internal or localhost-only resources and inject the response content into the model prompt context. This leads to exposure of sensitive internal information to unauthorized actors. The vulnerability does not impact integrity or availability but has a high confidentiality impact. Exploitation requires local access or user interaction (UI:R).
Mitigation Recommendations
A fix is available in PraisonAI version 4.6.40 and praisonaiagents version 1.6.40. Users should upgrade to these versions or later to remediate this vulnerability. Until upgraded, avoid processing untrusted prompt text containing '@url:' mentions or implement network restrictions to prevent unauthorized internal HTTP requests.
CVE-2026-47395: CWE-200: Exposure of Sensitive Information to an Unauthorized Actor in MervinPraison PraisonAI
Description
PraisonAI is a multi-agent teams system. Prior to version 4.6.40 of PraisonAI, corresponding to version 1.6.40 of praisonaiagents, PraisonAI's direct-prompt CLI automatically expands `@url:` mentions in raw prompt text before agent execution begins. If a prompt contains `@url:<http-or-https-url>`, the CLI calls `MentionsParser.process(...)`. The `@url:` handler then performs a direct `urllib.request.urlopen()` request to the attacker-controlled URL and returns the response body. That response body is prepended to the final model prompt context. There is no loopback/private-address restriction, no metadata-service restriction, and no approval gate before the fetch. As a result, attacker-influenced prompt text can cause the operator's machine to fetch localhost-only HTTP resources and inject the response into model context. PraisonAI version 4.6.40 and praisonaiagents version 1.6.40 contain a fix.
CVSS v3.1
Score 5.5medium
Affected software
pkg:github/mervinpraison/PraisonAIpkg:github/mervinpraison/praisonaiagentsRun on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
Weaknesses
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
PraisonAI's direct-prompt CLI feature automatically processes '@url:' mentions in raw prompt text by performing an HTTP(S) request to the specified URL using urllib.request.urlopen(). There are no restrictions on the requested URL, allowing attacker-controlled prompts to cause the operator's machine to fetch localhost or other internal network resources. The fetched response body is then prepended to the model prompt context, exposing sensitive internal information to unauthorized actors. This vulnerability is identified as CWE-200 (Exposure of Sensitive Information). Versions prior to 4.6.40 of PraisonAI and 1.6.40 of praisonaiagents are affected. The issue is fixed starting from these versions.
Potential Impact
An attacker who can influence prompt text can cause the operator's machine to make HTTP requests to internal or localhost-only resources and inject the response content into the model prompt context. This leads to exposure of sensitive internal information to unauthorized actors. The vulnerability does not impact integrity or availability but has a high confidentiality impact. Exploitation requires local access or user interaction (UI:R).
Mitigation Recommendations
A fix is available in PraisonAI version 4.6.40 and praisonaiagents version 1.6.40. Users should upgrade to these versions or later to remediate this vulnerability. Until upgraded, avoid processing untrusted prompt text containing '@url:' mentions or implement network restrictions to prevent unauthorized internal HTTP requests.
Technical Details
- Data Version
- 5.2
- Assigner Short Name
- GitHub_M
- Date Reserved
- 2026-05-19T19:22:45.729Z
- Cvss Version
- 3.1
- State
- PUBLISHED
- Remediation Level
- null
Threat ID: 6a5f96f32a4a8d59895fb31c
Added to database: 07/21/2026, 15:57:39 UTC
Last enriched: 07/21/2026, 16:12:50 UTC
Last updated: 07/21/2026, 19:42:45 UTC
Views: 9
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.