# VaultMesh Sentinel — Go-To-Market Battlecard (v1)

## What we are

VaultMesh Sentinel is the forensic continuity layer for autonomous infrastructure.

Sentinel makes systems **defensible after failure**, not merely secure during operation, by emitting offline-verifiable evidence of:
- what happened
- what was attempted and denied (Proof of Restraint)
- who/what had authority
- what corruption/tamper was detected

## Who we sell to (ICP)

Primary buyers:
- Space agencies & contractors (satellites, on-orbit servicing, lunar infrastructure)
- Critical IoT / OT operators (energy grids, pipelines, factories)
- Defense & national infrastructure vendors

Buyer personas:
- Program managers (mission liability)
- Security / safety leads (post-incident accountability)
- Compliance & legal (audit survival)
- Insurers (claim defensibility)

## The problem they already feel

- Automation is increasing faster than accountability.
- Systems operate offline, autonomous, and under coercion.
- After incidents, there is blame without proof; logs without integrity; narratives instead of evidence.

## Our wedge (why we win first)

**Proof of Restraint**

Sentinel produces auditable evidence not only of actions executed, but of actions **considered and safely denied**, with:
- denial reason (bounded + schematized)
- the exact operation that would have occurred (op + digest)
- any containment applied (scope narrowing)

## What Sentinel actually ships (v1)

- Action gating: intent → allow/deny → effect
- Append-only receipts + deterministic Merkle roots
- ShadowReceipts on denial (no silent drops)
- Corruption/tamper receipts and degraded-mode containment (authority can only narrow)
- Offline export bundles (seals) + offline verifier
- Archaeology drill as onboarding requirement

## The one-line pitch

“VaultMesh Sentinel is the black box recorder for autonomous infrastructure — it proves what happened, what was denied, and why, even years after failure.”

## Why now

- Automation is unavoidable (space latency, industrial scale)
- Regulation is tightening (NIS2 / CRA pressures)
- Insurance is demanding evidence, not promises
- Incidents are becoming political and international, not technical

## Competitive landscape (why others lose)

| Competitor type | Why they fail |
|---|---|
| SIEM / logging | Logs can be deleted, forged, coerced, or re-framed |
| Cloud governance | Assumes connectivity and a trusted control plane |
| Blockchains | Assumes liveness/consensus and pushes complexity into ops |
| Safety systems | Enforce rules but don’t prove restraint |
| Dashboards | Disappear after the incident |

Sentinel assumes the incident already happened.

## Proof artifacts (what we can hand an auditor)

Typical export bundle contains:
- `ROOT.current.txt` (root + seq + timestamp + algorithm identifiers)
- `receipts.jsonl` or a SQLite export covering the range
- `seal.json` (bundle metadata + ranges + root commitments)
- `integrity.json` (hashes of included files)
- `verifier_manifest.json` (expected tool versions/checksums)

## Pricing anchors (not promises)

Deployment licensing:
- Space / defense: $250k – $5M per system
- Critical IoT / OT: $50k – $500k per site

Recurring:
- Long-term support & verification tooling
- Compliance & evidence export packages

## First killer demo (closes deals)

**“The Black Box That Refused”**
1. Autonomous system runs offline.
2. Unsafe command is issued.
3. Sentinel denies it (ShadowReceipt emitted).
4. System continues safely.
5. Later, an auditor receives a proof bundle and verifies it offline.

Outcome: clear authority trail, provable restraint, zero ambiguity.

## Expansion path

1. Start as single-sovereign Sentinel (isolation-correct)
2. Add continuous invariant verification + drift containment
3. Optional federation for cross-witnessing (witness augmentation, not correctness)
4. Become a recognized evidence standard for autonomous operations