/

CVE-2026-90691

Severity:

HIGH RISK

(8.3)

Path Traversal (CWE-22) in HexStrike AI Files Endpoint

CVE-2026-90691

Deferred

Vulnerability type

Category:

Path Traversal

CWE:

CWE-22

Impact:

Unauthorized File Access or Modification

Severity:

HIGH RISK

(8.3)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

A path traversal vulnerability in HexStrike AI's API Files Endpoint allows a remotely connected attacker to influence the filename supplied to file operations. If the endpoint accepts insufficiently constrained path values, an attacker may access or manipulate files outside the intended directory.

Affected context

The vulnerability occurs in the FileOperationsManager function within hexstrike_server.py, where the API Files Endpoint processes a filename argument. Risk is present when attacker-controlled filename values are used in filesystem operations without reliably restricting them to an approved base directory. The remotely reachable endpoint and publicly disclosed exploit information may increase exposure for affected deployments.

Exploit preconditions

  • Attack surface: The remotely accessible API Files Endpoint - Required access: Network access to the endpoint; authentication

requirements depend on deployment configuration

  • Trigger point: Processing of the filename argument by FileOperationsManager - Why exploitation is possible: Insufficient path-boundary enforcement may

allow a supplied filename to resolve outside the endpoint's intended

filesystem area

Why this happens

The underlying bug is improper limitation of a user-controlled path during file handling. When an application combines or resolves an externally supplied filename without canonicalizing it and verifying that the result remains within an allowed directory, traversal sequences or equivalent path representations can escape the intended location. This behavior maps to CWE-22, Improper Limitation of a Pathname to a Restricted Directory.

Potential impact

  • An attacker may read sensitive files accessible to the service account.
  • An attacker could modify, overwrite, or delete files within the process's permissions.
  • Application configuration, credentials, or operational data may be exposed.
  • The service may experience data integrity loss or availability issues.
  • Compromise of accessible files could support further attacks.

General mitigation guidance

Mitigation generally focuses on reducing trust in filename input and enforcing a strong filesystem boundary. Secure patterns include accepting logical file identifiers rather than raw paths, resolving candidate paths against a fixed permitted directory, and rejecting any result that falls outside that directory. Applications can use allowlisted names, platform-aware path handling, and consistent authorization checks before every file operation. Sensitive files should be kept outside service-accessible locations, and the service should run with the minimum filesystem privileges required. Error responses should avoid disclosing filesystem details.

How teams detect issues like this

Detection approaches include static analysis for filesystem APIs that consume request parameters, data-flow review from API filename fields to open, read, write, rename, or delete operations, and code review checks for missing canonicalization and directory-boundary validation. Runtime monitoring can alert on requests containing traversal indicators, unusual path representations, or access attempts outside approved directories. API gateway policies can restrict filename formats and enforce authorization before file operations. Automated security review tools may help identify these patterns, but findings should be validated against deployment-specific path handling.

Sources

SourceLink
github.comgithub.com/0x4m4/hexstrike-ai/issues/135
github.comgithub.com/0x4m4/hexstrike-ai/issues/225
github.comgithub.com/0x4m4/hexstrike-ai/pull/159
vuldb.comvuldb.com/cve/CVE-2026-90691
vuldb.comvuldb.com/submit/914893
vuldb.comvuldb.com/vuln/403224
vuldb.comvuldb.com/vuln/403224/cti

Frequently asked

questions

Yes. The issue is associated with a remotely reachable API endpoint, although

successful exploitation still depends on network exposure and any

authentication or authorization controls in place.

An attacker generally needs access to the API Files Endpoint and must be able

to influence its filename argument. The exact requirement depends on whether

the deployment protects the endpoint with authentication and authorization.

Treat filenames as untrusted input, prefer allowlisted logical identifiers,

and resolve paths only within a fixed permitted directory. Verify the

canonical result before performing any file operation.

Use centralized path-validation utilities, apply least-privilege filesystem

permissions, review all request-to-filesystem data flows, and enforce

filename and authorization policies at the application and API gateway

layers.

Find out what's already
exploitable in your codebase.