JetBrains Junie
CVE-2025-58335
HIGH
Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessReach Junie agent in target IDE
-
ExploitCraft input invoking search_project
-
ExecutionFunction returns out-of-scope project data
-
ImpactSensitive project information disclosed
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A06:2021 - Vulnerable and Outdated Components
ASVS 5.0
V3 - Web Frontend Security Product UI inconsistency
|
| Exploitation Conditions | Requires interaction with the JetBrains Junie AI coding-agent feature and specifically its search_project function; the target must be running an affected Junie build (before 252.284.66 / 251.284.66 / 243.284.66 and the related 252/251/243.284.5x sub-builds) inside a JetBrains IDE. Exploitation is limited to information disclosure (confidentiality only - no code execution, modification, or DoS). The published AV:N/PR:N rating implies no authentication, but because Junie is an IDE-embedded assistant, realistic exploitation likely depends on the attacker being able to supply input the agent processes (directly as an IDE user or via injected/untrusted content); the advisory does not confirm a remote unauthenticated path against a default install. |
| Impact | Signals are consistently LOW-real-risk despite a 7.5 'High' base score. The NVD CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N) is confidentiality-only with no integrity or availability impact, and EPSS is 0.00% (0th percentile) - the lowest possible exploitation-probability signal. There is no CISA KEV listing and no public exploit identified, so exploitation status is 'no public exploit identified at time of analysis.' A key tension exists: the CVSS vector claims a network vector with no privileges (AV:N/PR:N), yet Junie is an AI coding-agent plugin embedded in JetBrains desktop IDEs, which normally implies local operator context - the AV:N rating is only coherent if the search_project function is reachable via untrusted/remote content processed by the agent (e.g. prompt injection), which the sparse advisory does not confirm. Verify the delivery path with JetBrains. The CWE-356 mapping (UI does not warn user of unsafe actions) is a poor fit for an information-disclosure flaw, for which CWE-200 would be canonical. Net: no SSVC 'act' urgency - 0% EPSS, no KEV, no POC, confidentiality-only, narrow product footprint make this a routine-patch item, not a priority. |
| Exploit ScenarioDerived | An attacker induces the Junie AI agent to invoke its search_project function in a way that returns file contents or project data outside the intended access scope, disclosing source code, secrets, or configuration to a party that should not see it. No public proof-of-concept exists and the exact trigger (direct user action vs. injected instruction from processed content) is not documented in the advisory. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv2 – Sep 26, 2026
Information disclosure in JetBrains Junie, the AI coding agent embedded in JetBrains IDEs, allows the assistant's search_project function to surface data beyond its intended scope on affected builds released prior to 252.284.66, 251.284.66, and 243.284.66 (with the 252/251/243.284.61, .284.54, and .284.50 sub-builds also listed as vulnerable). The impact is confidentiality-only - no code execution, data modification, or denial of service - and the published 7.5 'High' rating reflects an AV:N/PR:N/UI:N vector that independent review considers overstated, since realistic exploitation depends on a local, low-privilege actor interacting with the agent or supplying content it processes. Exploitation risk signals are weak across the board: EPSS is 0.00% (0th percentile), the CVE is not in CISA KEV, and no public exploit code has been identified at time of analysis.
Earlier versions of this analysis (1)
In JetBrains Junie before 252.284.66, 251.284.66, 243.284.66, 252.284.61, 251.284.61, 243.284.61, 252.284.50, 252.284.54, 251.284.54, 251.284.50, 243.284.54, 243.284.50 information disclosure was. Rated medium severity (CVSS 5.5), this vulnerability is no authentication required. No vendor patch available.
RemediationAI
Vendor-released patch: upgrade Junie to build 252.284.66 (2025.2 line), 251.284.66 (2025.1 line), or 243.284.66 (2024.3 line), or any later build, per the JetBrains fixed-issues advisory at https://www.jetbrains.com/privacy-security/issues-fixed/; users on the intermediate afflicted builds 284.61, 284.54, or 284.50 in those branches must move to 284.66 or newer. Where immediate upgrading is not possible, compensating controls include disabling or restricting the Junie agent's search_project tool (or the agent entirely) within IDE settings - accepting the loss of AI-assisted code search functionality - and limiting project indexing scope so the agent never reads unrelated directories or sensitive configuration files on the developer workstation. Because the independent assessment places exploitation at local access with user interaction rather than a remote unauthenticated path, the strongest practical mitigation is operational: avoid feeding untrusted or third-party content into Junie sessions and treat agent output as potentially containing off-scope data until patched. No public exploit code has been identified, and EPSS is 0.00%, so there is no basis for emergency out-of-band measures beyond the vendor patch.
Affects your stack
Find out whether the stacks and servers you run are on an affected version.
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
Junie is JetBrains' LLM-based coding agent that runs inside JetBrains IDEs (IntelliJ IDEA and siblings) and can invoke built-in tools such as search_project to index and query the user's codebase and project context. The flaw is mapped to CWE-356 (Product UI does not Warn User of Unsafe Actions): the agent's tool-invocation path in the affected builds executes search_project operations without adequate user warning or guardrails, so responses can include information the user never intended to expose. CPE entries (cpe:2.3:a:jetbrains:junie) confirm Junie itself is the affected component rather than the underlying IDE platform, with vulnerable builds spanning the 243.x (2024.3), 251.x (2025.1), and 252.x (2025.2) release lines. Notably, the published CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N vector implies an unauthenticated network-reachable path, whereas the independent assessment re-scores it as CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N, i.e. local access with low privileges; because the two vectors conflict, authentication requirements for real-world exploitation should be treated as not confirmed from available data.
Know when Junie gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N
Junie runs inside a local IDE session driven by an authenticated developer user, so AV:L and PR:L are more honest than NVD's AV:N/PR:N; impact is confidentiality-only.
Primary rating from NVD.
CVSS VectorNVD
Lifecycle Timeline
4DescriptionNVD
In JetBrains Junie before 252.284.66, 251.284.66, 243.284.66, 252.284.61, 251.284.61, 243.284.61, 252.284.50, 252.284.54, 251.284.54, 251.284.50, 243.284.54, 243.284.50 information disclosure was possible via search_project function
Executive ViewAI
JetBrains Junie, the AI coding agent embedded in JetBrains IDEs, has an information disclosure vulnerability that could expose sensitive project data to unauthorized users. While rated High severity, the risk is tempered by weak exploitation signals-no public exploit identified at time of analysis, not in CISA KEV, and EPSS 0.00%-and realistic exploitation requires a local, low-privilege actor. Leadership should prioritize protecting development environments from potential confidentiality breaches, as exposed intellectual property or credentials could have downstream business consequences.
No vendor-released patch identified at time of analysis. Interim mitigations: restrict Junie's access to sensitive data, disable search_project if possible, monitor for anomalous data access, and apply vendor patch when released.
Threat Intelligence Overview
NVD
CVE.org
EPSS
VulDB
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.
Affected Products (CPE)NVD
Command execution in JetBrains Junie before version 252.549.29 allows local attackers to execute arbitrary commands by c
In JetBrains Junie before 252.284.66, 251.284.66, 243.284.66, 252.284.61, 251.284.61, 243.284.61, 252.284.50, 252.284.54
GitHub PoC
Exploit-DB