high Cve 2026 109437 · AI Security

Cognition Agent Server Before 0.13.0 Lets Anyone on the Network Run Shell Commands (CVE-2026-109437)

Data graphic: editorial cover for CVE-2026-109437 in the Cognition agent server from CognicellAI: a huge 9.8 CVSS 3.1 Critical score beside a NO AUTH stamp, with stats of 9.2 under CVSS 4.0, no API authentication, and the fix in 0.13.0.
NS

Identity security analyst · Published Oct 31, 2026, 2:37 AM EDT

CognicellAI's self-hosted agent server shipped an unauthenticated API and a host-local shell backend by default. CVSS 9.8; upgrade to 0.13.0 and add auth.

Operators of Cognition, the self-hosted AI agent server from CognicellAI, should upgrade to 0.13.0. Every earlier version defaults to running agent commands directly on the host and exposes a REST/SSE API with no authentication. Anyone who can reach the server can get it to execute shell commands as the server process. The flaw is tracked as CVE-2026-109437, a VulnCheck-assigned record published on 19 October 2026.

What is affected

All Cognition releases before 0.13.0. Cognition is built on LangChain's Deep Agents framework and is meant to be run by teams on their own infrastructure, so exposure depends on who can reach the API. Any deployment reachable from an untrusted network is at risk, and so is one reachable only from a shared internal network.

How it works

Two defaults combine. The agent sandbox backend defaults to local execution, so model-directed commands run on the host, and the REST/SSE API has no authentication. An attacker creates a session, then posts a message that makes the agent run attacker-chosen shell commands. Lasso Security, which describes the flaw as an ungated shell tool, adds that credentials the backend passes through, such as GH_TOKEN, are reachable by the same route. It also says session scoping is off by default and relies on self-asserted headers. No user interaction is needed.

The CVE record lists CWE-1188 (insecure default initialization) and CWE-306 (missing authentication for a critical function). It scores 9.2 under CVSS 4.0 and 9.8 under CVSS 3.1, both Critical. The CISA ADP entry rates exploitation as PoC, automatable as No, and technical impact as Total. Lasso lists no known exploit. The researchers credited are Noy Pearl, Michael Sromin and Ophir Dror.

Data graphic: CVE-2026-109437 in the Cognition agent server, rated CVSS 9.8 Critical, drawn as a top-to-bottom chain: a network-reachable instance reaches an unauthenticated REST/SSE API, a session is created, the agent shell tool runs shell commands on the host, and credentials such as GH_TOKEN are exposed.

One unauthenticated API call to a reachable Cognition server ends in shell commands on the host and exposed credentials such as GH_TOKEN.

The fix

According to the CVE record, 0.13.0 refuses the local sandbox backend unless COGNITION_ALLOW_UNSAFE_LOCAL_EXECUTION is set to true, and the default is false. The release notes, dated 27 July 2026 in the changelog and published 28 July, describe the change more generally: strict production settings reject host-local execution and sandbox fallback, and exceptions need explicit development flags. They do not name that variable, and they describe no change to authentication. They put tenant authorization with the builder or operator.

The upgrade is not drop-in. In 0.13.0:

  • The implicit default agent is gone, so agents must be provisioned explicitly.
  • POST /sessions now requires an agent_name.
  • Writes must be drained before the migration, because mixed v0.12 and v0.13 writers are unsupported.
  • cognition db upgrade and cognition db current ship in the release image.

What defenders should do

  1. Upgrade to 0.13.0 or later, following the migration notes above.
  2. Do not set COGNITION_ALLOW_UNSAFE_LOCAL_EXECUTION=true on anything reachable by others.
  3. Put the API behind your own authentication, such as a reverse proxy or network policy, because the release notes describe no built-in authentication change.
  4. On instances that ran an affected version while reachable, review session and run history for unexpected sessions and shell commands, and rotate credentials the process could read, including GH_TOKEN.

Sources

Keep reading

All latest →
  1. elevatedAI SecurityClaude Code Before 2.1.251 Let Sandboxed Code Overwrite Files Outside the Workspace on macOS (CVE-2026-88968)3 min
  2. highAI SecurityFlowise and n8n Code Nodes Escape Their Sandboxes to Run OS Commands, and Flowise Has No Fix5 min
  3. elevatedAI SecurityRogue Relays and Prompt Adware Are Rewriting Codex and Claude Code Configs: What Symantec and Check Point Found5 min
  4. highAI SecurityMandiant: LLMjacking Bills Hit $100,000 a Day as Stolen Cloud Model Access Is Resold for $60 a Month5 min
  5. highAI SecurityEMBRACE: Embedded Browsers in Cursor, Copilot, Claude Desktop, Codex and VS Code Leak Files and Run Code4 min
  6. elevatedAI Securityllama.cpp CVE-2026-116816: A Crafted GGUF File Corrupts the Heap on Load, Fixed Since b102614 min