Security fixes are applied to the latest released version of braincell. We do not backport fixes to older minor versions — please upgrade before reporting an issue.

Please do not report security vulnerabilities through public GitHub issues, discussions, or pull requests. Doing so discloses the problem to everyone before a fix is available.

Instead, use one of these private channels:

1. GitHub private vulnerability reporting (preferred) — go to the Security tab and click Report a vulnerability. This keeps the report visible only to the maintainers until an advisory is published.
2. Email — chao.brain@qq.com.

Please include as much of the following as you can:

• The type of issue and the affected component or module.
• Full paths of the source files involved, and the version/commit affected.
• Step-by-step instructions to reproduce, ideally a minimal proof of concept. If the issue involves a morphology file, attach the smallest file that triggers it.
• Your environment: braincell, jax/jaxlib, brainstate, and Python versions; OS; and platform (CPU / GPU / TPU).
• The impact you believe the issue has, including how an attacker might exploit it.

These timelines may extend when maintainers are away, particularly around the end of the year. After the initial reply we will keep you informed of progress toward a fix and a public announcement, and may ask for additional information or guidance.

We support coordinated disclosure: we ask that you give us a reasonable opportunity to release a fix before publishing details. With your permission, we will credit you in the resulting security advisory.

braincell parses files and fetches data that a user may not have authored. Bugs that let such input affect confidentiality, integrity, or availability beyond what the documented API allows are in scope. In particular:

• Morphology parsing. Morphology.from_swc / Morphology.from_asc and the readers under braincell/io/ consume SWC and Neurolucida ASC files, which are routinely downloaded from public archives. Memory-safety or resource-exhaustion defects driven by file content — unbounded allocation, uncontrolled recursion, or a crash that escapes as something other than a clean exception — are worth reporting.
• Checkpoint loading. braincell/io/checkpoint.py reads .npz archives. It loads them with allow_pickle=False, so arbitrary-object deserialization is deliberately disabled; a way to reach a pickle-loading path anyway, or to escape the intended file set during extraction, would be a vulnerability.
• The NeuroMorpho.Org client. braincell.io.neuromorpho and the braincell-neuromorpho console script make network requests and write files to disk. Path traversal via server-supplied names, or transport weaknesses that let a network attacker substitute content, are in scope.

• braincell executes the models you give it. Model definitions, channel classes, and mechanism parameters are ordinary Python: constructing a Cell from attacker-supplied Python is arbitrary code execution, and is not a vulnerability. Treat model code with exactly the same trust you give any other code you import.
• Numerical disagreement is not a vulnerability. A simulation that diverges from NEURON, an integrator that loses accuracy on a stiff channel, or a solver that fails to converge is a correctness bug — please open a normal GitHub issue.
• Non-public API surface. Per section 6 of TODO.md, only the documented public API is a stable surface; anything else is internal and may change without deprecation. Reports that depend on reaching into private modules will generally be treated as ordinary bugs.
• Resource use under a legitimately large model. Simulating a very large cell is expected to consume a lot of memory. This is only a security matter when a small input causes disproportionate consumption.

Vulnerabilities in third-party dependencies (JAX, jaxlib, brainstate, brainunit, brainevent, braintools, brainpy, NumPy, SciPy, PyVista) are also out of scope. Report those to their respective maintainers — though we appreciate a heads-up if braincell is affected, so we can adjust our version constraints.

Security bugs in third-party modules should be reported to their respective maintainers. If the issue reaches users through braincell, please tell us too.