# Security — bild

> How bild works inside client systems: your stack, your guardrails, named approvals, least-privilege access and a record of every run.

Markdown rendering of https://bild.systems/security — the HTML version is canonical. Contact: humans@bild.systems

# Security.

Security is part of the method, not a page on the website. But here's the shape of it.

## How we work with client systems

bild builds workflows inside a client’s existing systems and security boundaries. Access is least-privilege and named, every action with a consequence passes a human checkpoint, and every run leaves a record your team can audit. Nothing we build is fully autonomous, and nothing leaves the room it already lives in.

- Workflows are built inside your existing stack and its security boundaries — your access controls, your data residency, your guardrails. Nothing leaves the room it already lives in.
- Access is least-privilege and named: we ask for what a workflow needs, not admin on everything.
- Every workflow keeps a record of what ran, what was drafted, who approved it and what was flagged — auditable by your team at any time.
- Nothing we build is fully autonomous. Actions with consequences pass through a named human checkpoint.

## This website

- Served over HTTPS from Cloudflare. It's a static site: no accounts, no sessions, no stored personal data.
- The enquiry form sends us an email and keeps nothing.

## Found something?

- If you've found a security issue on this site, tell us at humans@bild.systems. We'll reply, fix it, and say what changed.

Questions about any of this: humans@bild.systems.

Last updated September 2026.