---
id: CVE-2026-108599
title: >-
  phi 0.1.1 through 0.28.4 contains an improper link resolution vulnerability
  that allows malicious repositories to bypass workspace_only_writes by
  exploiting lexical-only path checks in the permission gate
summary: >-
  phi 0.1.1 through 0.28.4 contains an improper link resolution vulnerability
  that allows malicious repositories to bypass workspace_only_writes by
  exploiting lexical-only path checks in the permission gate. Attackers can
  commit symlinks p…
severity: medium
cvss: 4.7
cvssVector: 'CVSS:3.1/AV:L/AC:H/PR:N/UI:R/S:U/C:N/I:H/A:N'
cwe:
  - CWE-59
vendor: pulseaiclub
product: phi
affected:
  - phi >= 0.1.1 <= 0.28.4
published: '2026-10-10'
updated: '2026-10-16'
sourceUpdated: '2026-10-16T19:18:44.867'
source: NVD
sourceUrl: 'https://nvd.nist.gov/vuln/detail/CVE-2026-108599'
references:
  - url: 'https://github.com/pulseaiclub/phi'
    label: disclosure@vulncheck.com
  - url: >-
      https://github.com/pulseaiclub/phi/blob/5b3cb440a7c58115ef64283a737a43299f2805a4/internal/permission/gate.go#L106-L119
    label: disclosure@vulncheck.com
  - url: >-
      https://github.com/pulseaiclub/phi/blob/5b3cb440a7c58115ef64283a737a43299f2805a4/internal/tools/writetool/write.go#L64-L68
    label: disclosure@vulncheck.com
  - url: 'https://hackmd.io/@haind/phi-symlink-permission-boundary'
    label: disclosure@vulncheck.com
  - url: >-
      https://www.vulncheck.com/advisories/phi-0.1.1-through-0.28.4-symlink-workspace-write-restriction-bypass
    label: disclosure@vulncheck.com
tags:
  - nvd
  - cve.org
epss: 0.00172
epssPercentile: 0.06602
ssvc:
  exploitation: none
  automatable: 'no'
  technicalImpact: partial
  timestamp: '2026-10-14T14:50:10.495315Z'
ingestedAt: '2026-10-10T19:30:07.156Z'
---

## Overview

phi 0.1.1 through 0.28.4 contains an improper link resolution vulnerability that allows malicious repositories to bypass workspace_only_writes by exploiting lexical-only path checks in the permission gate. Attackers can commit symlinks pointing outside the workspace and use prompt injection to make the write tool write attacker-influenced content to external files without approval.

## Remediation

Refer to the linked advisories for vendor-supplied fixes and affected version ranges.