Database/AI/ML frameworks & serving
Obot: EOF line in a file-type MCP server config value breaks out of the init script and runs commands
Impact
Obot writes file-type MCP server configuration values into a heredoc in the init container script it generates. A value that contains an EOF line ends the heredoc early, so whatever follows runs as shell commands in that init container. From there the attacker can change the shared MCP server files volume, which other MCP servers and their users rely on, and can reach other services on Obot's Docker network. On a shared agent platform this turns any user who can configure an MCP server into code execution inside the platform's network.
Who can reach it
An authenticated Obot user who can create or edit a file-type MCP server configuration. Reachable over the network through Obot's normal interface.
What to do
Upgrade Obot to 0.27.0 or later (versions 0.9.0 through 0.26.x are affected) and restart the service. Review existing file-type MCP server configurations for stray EOF lines and check the MCP server files volume for unexpected changes. The advisory does not list a workaround.
References
Related entries
- BentoML OpenLLM 0.6.30 (async_run_command in src/openllm/common.py): A model repository directory name flows unescapedCVE-2026-15035 · BentoML OpenLLM 0.6.30 (async_run_command in src/openllm/common.py)Medium
- RAGFlow: unvalidated base_url in set_api_key lets authenticated users make server-side requestsCVE-2026-49951 · RAGFlow (POST /v1/llm/set_api_key base_url handling)Medium
- JupyterHub: unauthenticated logins write unbounded usernames to the log, exhausting storageCVE-2026-54338 · JupyterHub form-based login authenticators (failed-login logging)Medium
- vLLM: malformed JSON to the OpenAI-compatible endpoints returns server paths and versionsCVE-2026-73555 · vLLM OpenAI-compatible API server (validation_exception_handler, sanitize_message)Medium
- vLLM: attacker-supplied structured-output regex pins a CPU core and stalls the engine pathCVE-2026-73556 · vLLM structured outputs, lm-format-enforcer backend (structured_outputs.regex)Medium
- vLLM: integer overflow in the activation CUDA kernel leaks another batched request's outputCVE-2026-73558 · vLLM CUDA activation kernels (act_and_mul_kernel, activation_kernels.cu)Medium
This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.