From our desk to your stack

Sanity Check — external

Point-in-time public surface review: TLS, email auth, headers, exposed panels, cert hygiene. Plain-English report you own either way.

• Non-intrusive — no exploitation, no social engineering
• Findings ranked with fix order
• Soft path into a Sprint if you want us to drive fixes

Remote + access review

Sanity Check plus a short, read-only pass inside — identity, backup jobs, firewall exports, admin screenshare — once you grant revocable access.

• Time-boxed VPN / Tailscale / jump host
• Still fixed-scope; not a free-for-all pentest

Hardening Sprint — remote

Focused multi-week push against the fix list: harden what matters, document what changed, leave you with a cleaner baseline.

• Works over secure remote access
• Agent Hardening variant for shops already running AI tools

Agent deploy & shadow-AI containment

Scoped agent installs, permission gates, revoke drills, and policies so staff AI use stops being a silent leak path.

• Interview-style intake (not a 40-field form)
• “Notebook you can read” — transparent memory, not a black box

Policy, process, tabletops

Offboarding checklists, acceptable-use / AI policy that fits a small shop, incident and change lite, tabletop exercises over video.

• Built from real programs — stripped to SMB size
• Backup questions and restore-drill design (we don’t surprise-restore prod on a cheap audit)

Guardian — remote watch

Monthly external re-scan and plain-English delta. Optional log ship later for deeper digests. Quiet retainer, not a NOC cosplay.