If you believe you have found a security vulnerability in Muninn, please report it privately rather than opening a public issue.

Preferred: use GitHub's private vulnerability reporting on this repository.

Alternative: email the maintainer listed in the repository's GitHub profile, with subject line [muninn-security].

Please include:

• A description of the issue.
• Steps to reproduce.
• The affected version (git SHA or release tag).
• The impact you believe it has.
• Any suggested mitigation, if known.

You will receive an acknowledgement within 7 days. We aim to publish a fix or mitigation within 30 days of a confirmed report, depending on severity.

Muninn is a local-first research infrastructure project. The realistic threat surface in the MVP is small but not zero. In-scope reports include:

• Vulnerabilities in the ingestion API or query API that allow unauthorized access, data corruption, or remote code execution.
• Vulnerabilities in event validation that allow malformed events to corrupt the event log or break deterministic replay.
• Vulnerabilities in dependencies that materially affect Muninn's runtime.
• Configuration defaults that expose sensitive surfaces unintentionally.

Out of scope:

• Findings that require an attacker who already has local shell access to the host running Muninn.
• Issues in third-party services Muninn integrates with (report those to the relevant project).
• Denial-of-service via unbounded event submission to a local development deployment — this is documented and not a security issue at the MVP scope.
• Speculative reports without reproduction steps.

Muninn is pre-1.0. Only main is currently supported. Once tagged releases exist, the latest minor version will receive security fixes.

We follow standard coordinated disclosure: a private fix lands first, then a public advisory is published with credit to the reporter (unless anonymity is requested).

Muninn is not hardened for adversarial environments. If you deploy it outside localhost:

• Put authenticated reverse-proxy authentication in front of the query API and ingestion API.
• Never expose Redpanda, PostgreSQL, or MinIO ports to the public internet without authentication.
• Use TLS for any port exposed beyond the host.
• Store secrets in environment variables or a secret manager — never in committed configuration.

See also the upcoming docs/steering/SECURITY_MODEL.md for the broader threat model.