Family D — Agent Governance & Boundary Assurance
Every tool or connector an agent uses is a new trust boundary.
This is an independent review of where the agent's authority actually ends — not where policy assumed it did.
Opens your email app with the intake questions pre-filled.
Fit / not-fit
This fits if
- Agents connect to tools, servers, or third-party integrations.
- You do not yet have an independent boundary review.
- The connector surface is growing faster than the review process.
This is not a fit if
- You want a generic identity / access-control product.
- You expect a conventional penetration test as the deliverable.
- You have no agent or connector surface yet.
Offer and current pricing
| Offer | What it is | Price | Status |
|---|---|---|---|
| Grouped boundary and governance track | Vendor acceptance review, technical due diligence, connector-boundary review, and evidence-first governance workshop — scoped as one track. | No individual public list price yet | Scoped on request |
| Workflow admission / no-bypass review | Incubating module under this family. | No public price | Incubating |
Rather than invent four price points that do not exist, this page states the gap: fee is set per engagement once scope is confirmed by reply.
What the finding looks like
Illustrative only — no real client, no real finding, no protected methodology:
Of 6 tool boundaries tested, 4 held as configured; 2 allowed an action beyond the documented scope under a specific error-string injection — reported with the exact condition that triggered it, not a generic maturity score.
Limitations
After a fit reply: Statement of Work
A short SOW states the exact boundaries to be reviewed and the fee for that engagement. No public list price exists to quote here. Work starts after written acceptance.