CVE-2026-61536
HighCVSS 7.5Exploitation Probability (EPSS)
Low risk22th percentile - higher than 22% of all known CVEs
Summary
Banks prior to version 2.4.3 contains a remote code execution vulnerability in {% completion %} blocks. An attacker can supply a malicious Tool JSON object with an import_path field that is resolved via importlib.import_module() and getattr() without any allowlist. When the LLM invokes a tool with a matching name, arbitrary Python code such as os.system or subprocess.getoutput can be executed.
Risk Assessment
Risk of full compromise of the Banks-hosting process, potentially leading to loss of confidentiality, integrity, and availability.
Recommendation
Immediately update Banks to version 2.4.3 or later.
Other vulnerabilities in Banks
See all- CVE-2026-71492Medium
In Banks before version 2.4.5, DirectoryPromptRegistry.set() in src/banks/registries/directory.py interpolates attacker-controlled Prompt.name and Prompt.version values into a Path without canonicalization or containment validation. Relative traversal such as ../victim/foo and an absolute Prompt.name can escape or discard the configured registry root, while overwrite=True permits replacement of existing target files. The poisoned name is persisted in index.json and reconstructed by _load(), allowing the out-of-root path to survive later registry loads. An application that forwards request data into these fields can therefore write Prompt.raw bytes to attacker-chosen paths writable by the application process. This issue is fixed in version 2.4.5.
- CVE-2026-62663High
Banks generates meaningful LLM prompts using a simple template language. In versions prior to 2.4.4, all four media filters (image, audio, video, document) in banks accept untrusted user input as file paths via Path(value) and pass them directly to open(file_path, "rb") without any path sanitization, canonicalization, or directory restriction. An attacker who controls template variables passed to a banks Prompt can use path traversal (../) to read arbitrary files accessible to the Python process—including .env files, SSH keys, cloud credentials, source code, /etc/passwd, and /etc/shadow—with the content returned base64-encoded in the rendered prompt output, making exfiltration trivial.
- CVE-2026-44209High
Banks before version 2.4.2 uses unsandboxed jinja2.Environment() to render prompt templates, allowing Server-Side Template Injection (SSTI) leading to Remote Code Execution (RCE) on the host.
Original NVD description (English source)
Banks generates meaningful LLM prompts using a simple template language. In versions prior to 2.4.3, banks parses Tool JSON objects from the rendered body of {% completion %} blocks and later resolves their import_path field through importlib.import_module(...) + getattr(...) to obtain the callable that handles a tool call. There is no allowlist or sanitization on import_path, so any importable Python attribute (e.g. os.system, subprocess.getoutput) can be selected. When the LLM emits a tool_calls entry whose function.name matches the attacker-supplied tool name, the resolved callable is invoked with kwargs decoded from tool_call.function.arguments, yielding arbitrary code execution in the banks-hosting process. This is distinct from GHSA-gphh-9q3h-jgpp / CVE-2026-44209. That advisory was fixed in 2.4.2 by switching src/banks/env.py from Environment to SandboxedEnvironment. The fix does not touch src/banks/extensions/completion.py, and the unsafe import + getattr chain still executes on 2.4.2. The malicious Tool JSON is plain text in the rendered template body — it requires no Jinja attribute access, so the sandbox is irrelevant. This issue has been fixed in version 2.4.3.