About

Security for the people
shipping with AI.

Sigil checks the app you shipped, finds what AI builders leave exposed, and tells you how to fix it in plain English. No security team required.

The short version01
Blackbox first.
No source upload required.
ownerDNS verified before any request
read onlyno writes, no exploitation
sourceoptional, sandboxed, deleted after the scan
humanreviewed reports when you need one

Why we exist

AI made shipping easy. It did not make security automatic.

A founder can now ship a working app in a weekend. The same speed produces the same mistakes: a public database, a key in the browser bundle, an admin panel with no rate limit. The person who built it is often the only person who can fix it, and they are not a security engineer.

Sigil is built for that person. The report starts with what is exposed, who can reach it, and the exact fix. The scan is read only. The owner has to prove control of the domain before we send a request.

Principles

What we will and will not do

Verify

Owner first

DNS verification before an active scan. No drive-by scanning of apps you do not control.

Read

No exploitation

Read only checks. No writes, no brute force, no destructive payloads.

Explain

Plain English

The evidence, the impact, and the fix. Not a wall of CVSS scores.

Source

Optional by design

Blackbox is the default. Source is optional, sandboxed, and deleted after the scan.

Human

Accountable

A person reviews the report when you ask for a signed one.

Honest

Blocked is not clean

If a WAF or login wall stops the check, the report says inconclusive. It never becomes an A by default.

Built by

MischiefStack

Sigil is a MischiefStack project. We build small, sharp tools for people who ship.