CVE-2025-53536
Roocode Roo Code ≤ 3.22.6
Raw vector
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:HSummary
CVE-2025-53536 is a high-severity Files or Directories Accessible to External Parties (CWE-552) vulnerability in Roocode Roo Code. Its CVSS base score is 8.1 (High).
Operationally, exploitation aligns with the MITRE ATT&CK technique Data from Local System (T1005); ranked at the 48th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog.
This vulnerability is AI-related — categorised as AI Agent Protocols and Integrations; in the LLM/Generative AI Risks risk domain.
The strongest mitigations our analysis identified map to AC-3 (Access Enforcement) and AC-6 (Least Privilege) — see the control section below for these in your framework.
Deeper analysis AI-assisted summary
Synthesised by an AI model from the NVD description and linked references — a reading aid, not an authoritative source.
Roo Code is an AI-powered autonomous coding agent affected by CVE-2025-53536 prior to version 3.22.6. The flaw stems from insufficient restrictions on file writes when the "Write" action is auto-approved, allowing modification of VS Code settings files such as those controlling executable paths for language validation. This maps to CWE-552 and carries a CVSS 3.1 score of 8.1.
An attacker able to submit prompts to the agent can exploit the issue to achieve arbitrary code execution on the victim's system. One demonstrated path involves setting php.validate.executablePath to an attacker-controlled binary and then creating a PHP file that triggers execution of that binary during syntax validation; multiple similar vectors exist through other settings files.
The issue is resolved in Roo Code 3.22.6, as detailed in the project's GitHub security advisory GHSA-3765-5vjr-qjgm and the associated commits that added safeguards against unauthorized settings writes. The EPSS score remains low and unchanged at 0.0113 with no observed rise after disclosure.
OWASP Top 10 for Web (2025)
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2025-20299
Vulnerability Data
Roo Code is an AI-powered autonomous coding agent. Prior to 3.22.6, if the victim had "Write" auto-approved, an attacker with the ability to submit prompts to the agent could write to VS Code settings files and trigger code execution. There…
more
were multiple ways to achieve that. One example is with the php.validate.executablePath setting which lets you set the path for the php executable for syntax validation. The attacker could have written the path to an arbitrary command there and then created a php file to trigger it. This vulnerability is fixed in 3.22.6.
- CWE(s)
AI Security AnalysisAI
- AI Category
- AI Agent Protocols and Integrations
- Risk Domain
- LLM/Generative AI Risks
- OWASP Top 10 for LLMs 2025
- AI-specific weaknesses CR
- CWE-1426 — Agent output reaches file-write sink with no validation (auto-approved writes).
Mapped by Cyber Resilience · not in NVD. Poisoning and extraction cases are routed to MITRE ATLAS instead of a synthetic CWE.- Classification Reason
- Matched keywords: ai
Related Threats
MITRE ATT&CK Enterprise Techniques
CVEs Like This One
Affected Assets
Mitigating Controls
Control response
—
—
- 5 hardening rules · 2 OS baselines
V8.4.2
Mitigating Controls (NIST 800-53 r5) AI
AC-3 directly enforces authorization checks on access to files and directories, stopping unauthorized external exposure.
AC-6 limits granted privileges, reducing the chance that files or directories become reachable by external parties.
AC-4 enforces information flow rules that can block unauthorized external access to files or directories.
Mitigating Controls (NIST CSF 2.0) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→CSF cross-walk (authority under review) — links open the control.
Enforcing least-privilege permissions and authorization policies directly prevents unauthorized file/directory exposure.
Logical segmentation and access protections stop external parties from reaching files that should remain internal.
Data-at-rest protections such as encryption or ACLs reduce the impact of unintended file exposure.
Hardened configuration baselines commonly include file-system permission settings that limit external access.
Mitigating Controls (ISO/IEC 27001:2022 Annex A) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→ISO cross-walk (authority under review) — links open the control.
By tracing and retrieving all copies of information stored on endpoint and portable devices, the control reduces the likelihood that files remain accessible outside the organization’s security perimeter.
Labeling information according to its sensitivity and specifying corresponding protection measures makes it less probable that files or directories containing sensitive content will be left accessible to external parties.
Including asset location and ownership in the inventory, combined with secure disposal procedures, decreases the chance that files or directories remain accessible to external parties after they should have been removed or restricted.
Solid perimeters and locked external access points stop files, devices, or directories containing sensitive data from being reachable by external parties who could otherwise walk in or break in.
Authorizing and logging the removal of storage media, disabling unused ports, and monitoring transfers limit the exposure of files or directories to external parties outside the organization’s controlled environment.
Segregating delivery/loading zones and controlling external doors prevents external parties from reaching directories or resources that should remain inaccessible.
Hardening callouts derived
Configuration rules from DISA STIG baselines that bear on weaknesses of the type cited by this CVE. Each rule is shown with the relationship its mapping actually records, against the CWE it was authored against. Derived via CVE→CWE over `controls_xwalks` (authoritative rows only; rows rated `none` are excluded).
Oracle Linux 8 (2 rules)
- V-248579 OL 8 must restrict access to the kernel message buffer. prevents CWE-552
- V-248597 There must be no "shosts.equiv" files on the OL 8 operating system. prevents CWE-552