Incord Intelligent Language (inlang), by Incord AI, compiles a .az file into
a website and, optionally, a backend (accounts, data tables, chat). Security problems in the compiler, in the sites and servers
it generates, or in the CLI tools are taken seriously.
Do not open a public issue, pull request or discussion for a security problem.
Report it privately on GitHub: https://github.com/incordai/inlang/security/advisories/new ("Report a vulnerability" on the repository's Security tab). Only the maintainers can see the report.
Please include:
- what is affected: the compiler, a generated site, the generated backend, the
chat block, a CLI command (
az build,az dev,az mcp, …); - the
.azfile, configuration or request that shows the problem — the smallest one you can make; - what an attacker can do with it;
- the commit (
git rev-parse HEAD) you tested.
Never include real secrets, keys or other people's personal data in a report.
The maintainers confirm they have received the report, investigate it privately, and agree a disclosure date with you. The fix is developed in a private advisory, released, and then the advisory is published with credit to you unless you ask to stay anonymous.
az has not made a release yet. Security fixes go to the main branch; use the
latest main.
In scope: anything that lets an attacker run code, read data they should not,
bypass sign-in or a data table's allow rules, ship a secret in a built site,
weaken a generated site's Content-Security-Policy, or make the chat block
render content outside its allowed components.
Out of scope: a vulnerability in a site author's own custom code (raw
components, ejected words), and findings that need an attacker who can already
edit the project's files or the server's .env.
The checks that run on every change — secret scanning of every build, the generated sites' security policy, the scanners in CI — are described in docs/security-scanning.md and docs/USAGE.md §5.