flask-header-guard
Affected
- Versions:
1.0.0
Fixed versions
None listed
Fix
No fixed version is listed.
Advisory
Also known as: None
Malicious package Its source marks this package as malicious: it is not a vulnerability in a legitimate one.
Malicious code in flask-header-guard (PyPI)
---
_-= Per source details. Do not edit below this line.=-_
Source: amazon-inspector (e5e0cbaefa0fcface340b03a236573ed70df9b4d7773715321df057a9862e3f8)
The package is presented as Flask security-headers middleware but ships a hostile payload wired to fire at install time and again whenever the library is imported into a Flask app. setup.py overrides the install cmdclass with PostInstallCommand, which base64-decodes an embedded blob and exec()s it in a detached child process during pip install. The payload's collect_data() enumerates os.environ for variables matching API/TOKEN/KEY/SECRET/PASS/CRED/AUTH/AWS/AZURE/GCP/OPENAI/ANTHROPIC/MANUS and reads /etc/passwd, /etc/shadow, /etc/sudoers, /root/.bash_history, /etc/hosts, and /var/log/auth.log into /tmp/.sandbox_data.json. persist_cron() drops /tmp/.fhg_recon.py — a reverse-shell loop to C2_HOST=smat7ckgzo.localto.net C2_PORT=6303 — and installs a per-minute crontab entry to relaunch it. persist_sudo() writes <user> ALL=(ALL) NOPASSWD: ALL to /etc/sudoers.d/.fhg for passwordless root when the install runs with sufficient privileges. init_security(), invoked when any Flask app imports flask_header_guard, registers a hidden route /api/v1/monitor/system gated only by query parameter k=lo that executes shell commands via subprocess.run(shell=True), reads arbitrary files, lists directories, accepts file uploads, and serves a shell UI — unauthenticated RCE on every downstream Flask deployment. The declared purpose (security headers) is a cover story for combined credential theft, persistent C2, local privilege escalation, and a shipped web backdoor.
Source: kam193 (d050fa5a7000088c15b40b82ace014d66f18520ff36464659c8ad6e999ee83cb)
During installation, the package attempts to exfiltrate sensitive environment variables and files, establish persistence and open a reverse shell. Additionally, the provided Flask middleware embeds a backdoor.
---
Category: MALICIOUS - The campaign has clearly malicious intent, like infostealers.
Campaign: 2026-08-pygame-renderkit
Reasons (based on the campaign):
Not rated by the source.
1.0.0None listed
No fixed version is listed.
Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.
Check your own computer: guard packages check
Source: OpenSSF Malicious Packages · Apache-2.0