Security overview

I work through access you grant, and you can remove it in one step.

• Sign-in through your identity provider, with multi-factor authentication.
• Roles scoped to the work at hand. Production access only when a task needs it.
• Every access is logged in systems you control.

My own machines are Apple devices with biometric sign-in, full-disk encryption, and automatic screen lock.

Production data stays in your environment. I don't copy it to my machines or to any system I operate.

• Development and testing use synthetic or de-identified data.
• Sensitive fields can be encrypted per client while staying searchable.
• Audit events are chained together, so an edited or deleted record shows.

Nothing reaches production without a record.

• Every change goes through a pull request in your repository.
• Automated tests run before anything merges.
• Deploys run from your pipeline, so the history of what shipped, and when, lives with you.

You decide which outside services see your data, and every call is on record.

• Each AI or data provider is listed and approved by you before it's used.
• Every outside call leaves a receipt: what was sent, for which records, and what came back.
• Documents the system reads are wrapped in a marker that changes every time, so text hidden inside a file can't pose as an instruction.
• Where the evidence doesn't support an answer, AI features are built to decline rather than guess.

Clients on the Run and Build plans get outage response and security upkeep. That covers monitoring, patches, dependency updates, and verified backups.

Response expectations are written into each plan agreement, not promised on a web page.

Ending the relationship takes one step: remove my access. The code, data, documentation, and runbooks are already in your accounts, so there is nothing to hand back.