CVE-2026-54051 Network-AI has an an OS Command Injection issue

Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.9.1, the agent sandbox gates shell commands behind an allowlist SandboxPolicy.isCommandAllowed, which THREATMODEL.md calls the main control against a compromised agent Adversary 3.2. The allowlist glob-matches the who...

CVE-2026-54051

Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.9.1, the agent sandbox gates shell commands behind an allowlist SandboxPolicy.isCommandAllowed, which THREATMODEL.md calls the main control against a compromised agent Adversary 3.2. The allowlist glob-matches the who...

CVE-2026-54051

Network-AI (TypeScript/Node.js) prior to 5.9.1 is vulnerable: the agent sandbox gates shell commands via an allowlist, but ShellExecutor runs commands through /bin/sh -c. Wildcards like git *, npm *, or node * can match unintended strings (e.g., git status; …), enabling arbitrary code execution i...

CVE-2026-54051 Network-AI has an an OS Command Injection issue

Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.9.1, the agent sandbox gates shell commands behind an allowlist SandboxPolicy.isCommandAllowed, which THREATMODEL.md calls the main control against a compromised agent Adversary 3.2. The allowlist glob-matches the who...

CVE-2026-54051 Network-AI has an an OS Command Injection issue

Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.9.1, the agent sandbox gates shell commands behind an allowlist SandboxPolicy.isCommandAllowed, which THREATMODEL.md calls the main control against a compromised agent Adversary 3.2. The allowlist glob-matches the who...

CVE-2026-64623

Network-AI before 5.13.4 contains an improper cryptographic signature verification vulnerability in APSAdapter where the default local verifier accepts any non-empty string as valid. Unauthenticated attackers can submit forged APS delegation payloads with arbitrary scopes to bypass signature...

CVE-2026-64623 Network-AI before 5.13.4 Cryptographic Signature Verification Bypass

Network-AI before 5.13.4 contains an improper cryptographic signature verification vulnerability in APSAdapter where the default local verifier accepts any non-empty string as valid. Unauthenticated attackers can submit forged APS delegation payloads with arbitrary scopes to bypass signature...

CVE-2026-64623

Network-AI before 5.13.4 contains an improper cryptographic signature verification vulnerability in APSAdapter where the default local verifier accepts any non-empty string as valid. Unauthenticated attackers can submit forged APS delegation payloads with arbitrary scopes to bypass signature...

EUVD-2026-45938

Network-AI before 5.13.4 contains an improper cryptographic signature verification vulnerability in APSAdapter where the default local verifier accepts any non-empty string as valid. Unauthenticated attackers can submit forged APS delegation payloads with arbitrary scopes to bypass signature...

CVE-2026-64623 Network-AI before 5.13.4 Cryptographic Signature Verification Bypass

Network-AI before 5.13.4 contains an improper cryptographic signature verification vulnerability in APSAdapter where the default local verifier accepts any non-empty string as valid. Unauthenticated attackers can submit forged APS delegation payloads with arbitrary scopes to bypass signature...

PT-2026-61688

Name of the Vulnerable Software and Affected Versions Roo Code versions prior to 3.54.1 Description A command injection issue exists in the auto-approve execute feature. Attackers can bypass allowlist and denylist enforcement by nesting command substitutions within parameter expansion defaults. T...

PT-2026-61705

Name of the Vulnerable Software and Affected Versions GPT-SoVITS versions prior to 20250606v2pro Description An OS command injection flaw exists in webui.py where the ASR, slice, denoise, and uvr5 functions interpolate unsanitized Gradio textbox values directly into shell commands executed with...

CVE-2026-58195 Agentic-Flow: OS Command Injection in agentic-flow MCP server tools via unsanitized tool-parameter interpolation into execSync

Agentic-Flow is an AI agent orchestration platform. Prior to 2.0.14, agentic-flow MCP server tools in src/mcp/standalone-stdio.ts, src/mcp/fastmcp/servers/claude-flow-sdk.ts, src/mcp/fastmcp/servers/stdio-full.ts, src/mcp/fastmcp/servers/http-streaming-updated.ts,...

CVE-2026-58195

Agentic-Flow (MCP server tools) prior to version 2.0.14 is vulnerable to OS command injection. In affected code paths, tool parameters such as agent, task, name, language, and agentdb could be interpolated directly into shell command strings passed to execSync(), enabling arbitrary command execut...

Pheditor: Incomplete command sanitization in terminal feature allows RCE via pipe operator, backtick substitution, and newline injection

Summary The terminal feature in Pheditor uses an incomplete character blocklist to sanitize user-supplied commands before passing them to shellexec. After the fix for GHSA-9643-6xjp-vx57 which added $ to the blocklist, the characters | single pipe, backtick, and the newline byte 0x0A remain...

GHSA-WG4W-WR5Q-6VJC Pheditor: Incomplete command sanitization in terminal feature allows RCE via pipe operator, backtick substitution, and newline injection

Summary The terminal feature in Pheditor uses an incomplete character blocklist to sanitize user-supplied commands before passing them to shellexec. After the fix for GHSA-9643-6xjp-vx57 which added $ to the blocklist, the characters | single pipe, backtick, and the newline byte 0x0A remain...

Pheditor has an authenticated terminal command whitelist bypass

Summary Pheditor 2.0.4 has an authenticated terminal command whitelist bypass. The terminal feature checks whether the submitted command starts with one of the configured TERMINALCOMMANDS values, then passes the full command string to shellexec. Shell command substitution such as $ is not blocked...

GHSA-9643-6XJP-VX57 Pheditor has an authenticated terminal command whitelist bypass

Summary Pheditor 2.0.4 has an authenticated terminal command whitelist bypass. The terminal feature checks whether the submitted command starts with one of the configured TERMINALCOMMANDS values, then passes the full command string to shellexec. Shell command substitution such as $ is not blocked...

Malicious code in @cyberrant-rantai/rantai (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 41d1dbf6fb83f9a7c4c4b94ee40255e86910714400e249fec55b031e809c92bd runtime/localagent.py opens a Socket.IO client connection to a hardcoded server at https://app.cyberrant.org overridable via $SERVERURL and registers...

MAL-2026-10749 Malicious code in sync-grove (npm)

--- -= Per source details. Do not edit below this line.=- Source: amazon-inspector 65248cd6e2924b6485824bcf382441e39d5e09860bc1e849aee0d6d289d184fb Package.json declares "postinstall": "node setup.js", so setup.js runs automatically on npm install. The script assembles the strings shelljs and exe...