gcdRadar
archive/Patch now/AIR-2026-1857
closedclosed 10 Nov 2026Patch now2 Disclosed vulnerabilityAI

CVE-2026-113663: SandboxShellTool newline injection bypasses allowlist, enabling arbitrary command execution

SandboxShellTool

A newline injection flaw in SandboxShellTool's allowed and readonly modes lets an attacker embed a newline to break the command filter and execute arbitrary shell commands via sh -c. The vulnerability is tracked as CVE‑2026‑113663 and currently has no known exploitation.

why this severityDisclosed vulnerability with CVE, no confirmed exploitation.

disclosed CVENo confirmed exploitation, so this record is tracked as a disclosed vulnerability and not counted as an incident. It moves into the counts if CISA or the vendor confirms exploitation. Why

CVSS
8.8
EPSS
<0.01 · top 72% · 2026-11-13
fix status
Fix status unknown
affected
ag2 >= 0.13.2, < 1.1.1
autogen >= 0.13.2, <= 0.14.1
fixed in
ag2 1.1.1
first report
10 Nov 2026
last update
10 Nov 2026

What to do

Avoid using untrusted input with SandboxShellTool and monitor for a patched release.

Timeline · 1

  1. 10 Nov 2026CVE-2026-113663: AG2: Newline injection in SandboxShellTool bypasses allowed=[...] and readonly=True, enabling arbitrary command execution

Sources · 1

Advisories and databases

  1. 2026-11-10

Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.

Classification

AI role
Attack on AI
categories
Infrastructure vulnerability
severity
2 · Disclosed vulnerability
fix status
Fix status unknown
targets
Framework / tooling
owasp llm top 10
LLM03:2025 Supply Chain
mitre atlas
no technique for these categories
dates
happened 10 Nov 2026 · first reported 10 Nov 2026
origin
GitHub Advisory Database
Vulnerability record from NVD / the GitHub Advisory Database. This product uses the NVD API but is not endorsed or certified by the NVD.
record
confidence 0.95 · created 2026-11-10 16:11Z · updated 2026-11-10 16:11Z

Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned

Copy a brief

This record as a short Markdown brief, for a ticket, a chat or a coding agent.

the brief as text
# CVE-2026-113663: SandboxShellTool newline injection bypasses allowlist, enabling arbitrary command execution

https://gcdradar.vercel.app/incidents/cve-2026-113663-sandboxshelltool-newline-injection-bypasses · AIR-2026-1857 · closed 2026-11-10

- Lane: Patch now
- Severity: 2 (Disclosed vulnerability): Disclosed vulnerability with CVE, no confirmed exploitation.
- Fix status: Fix status unknown
- First reported: 2026-11-10
- AI role: Attack on AI
- Products: SandboxShellTool
- CVE ids: CVE-2026-113663
- CVSS: 8.8

## Summary

A newline injection flaw in SandboxShellTool's allowed and readonly modes lets an attacker embed a newline to break the command filter and execute arbitrary shell commands via sh -c. The vulnerability is tracked as CVE‑2026‑113663 and currently has no known exploitation.

## What to do

Avoid using untrusted input with SandboxShellTool and monitor for a patched release.

## Affected and fixed versions

- Affected: ag2 >= 0.13.2, < 1.1.1
- Affected: autogen >= 0.13.2, <= 0.14.1
- Fixed in: ag2 1.1.1

## Latest updates

- 2026-11-10: CVE-2026-113663: AG2: Newline injection in SandboxShellTool bypasses allowed=[...] and readonly=True, enabling arbitrary command execution (https://github.com/advisories/GHSA-2h4w-29cv-2cmv)

## Sources

- 2026-11-10 GitHub Advisory Database: CVE-2026-113663: AG2: Newline injection in SandboxShellTool bypasses allowed=[...] and readonly=True, enabling arbitrary command execution https://github.com/advisories/GHSA-2h4w-29cv-2cmv

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.

Related

  1. closed6.3
    Flowise CVE-2026-127214: Flowise server-side request forgery vulnerability

    Patch now · Disclosed · since 13 Nov 2026 · CVE-2026-127214

    AI
  2. closed6.9
    vLLM CVE-2026-126395: vLLM information disclosure vulnerability in sanitize_message()

    Patch now · Disclosed · Patched · since 13 Nov 2026 · CVE-2026-126395

    AI
  3. closed7.1
    vLLM CVE-2026-126399: vLLM resource exhaustion vulnerability via video frame limits bypass

    Patch now · Disclosed · Patched · since 13 Nov 2026 · CVE-2026-126399

    AI
  4. closed8.2
    vLLM CVE-2026-126398: vLLM reachable assertion vulnerability in GPUModelRunner._init_xdrope_positions

    Patch now · Disclosed · Patched · since 13 Nov 2026 · CVE-2026-126398

    AI
  5. closed6.9
    vLLM CVE-2026-126394: vLLM resource exhaustion vulnerability in OpenCVDynamicOpenPanguVideoBackend

    Patch now · Disclosed · Patched · since 13 Nov 2026 · CVE-2026-126394

    AI
  6. closed8.7
    vLLM CVE-2026-126397: vLLM quadratic CPU work vulnerability in Cohere2 Vision multimodal preprocessing

    Patch now · Disclosed · Patched · since 13 Nov 2026 · CVE-2026-126397

    AI