Skip to content

Security: leafmonkeylabs/salli-core

SECURITY.md

Security

Salli holds people's financial records, so we take reports seriously.

Please report vulnerabilities privately through GitHub's private vulnerability reporting rather than a public issue, or by email to security@salli.money. Include what you found, how to reproduce it, and the impact you expect.

What happens next

Step Within
We acknowledge your report 3 working days
We confirm it, or say why not, with a severity 10 working days
A fix released for a critical or high severity issue 30 days of confirming it
A fix released for anything else the next minor release

We keep you updated until it's fixed, credit you in the advisory unless you would rather we didn't, and publish a GitHub security advisory when the fix ships. Please give us 90 days, or until the fix is out if sooner, before you disclose. We won't pursue anyone who reports in good faith, keeps to their own account and data, and gives us that time.

Supported versions

Salli is before 1.0: only the latest release of each part gets security fixes.

Part Supported
The server image (ghcr.io/leafmonkeylabs/salli) and the console the latest server-v release
The salli CLI (@leafmonkeylabs/salli) the latest cli-v release
The SDK (@leafmonkeylabs/salli-sdk) the latest sdk-v release
main yes, as the next release

The threat model and controls are in docs/security.md; what personal data Salli keeps and who else sees it is in docs/privacy.md.

Running your own instance safely

  • Keep .env out of version control and readable only by you (salli-server setup writes it with mode 600).
  • There is no sign-up: only accounts the owner creates can sign in. Keep the secrets volume and your backups as private as the database.
  • Never set SALLI_INSECURE_DEV_AUTH outside local development.
  • If you expose the API beyond your machine, put it behind HTTPS (compose's tls profile) and set SALLI_PUBLIC_URL to its public URL. The console signs in only over HTTPS, or at localhost on the server's own machine.
  • docs/self-host.md says how passwords, sessions, refresh tokens (a reused one ends its sign-in), the console's headers and sign-in throttling work.

There aren't any published security advisories