Nexus Synergy Engagement Package — B01 · Irish Naval Service / Defence Forces RTI Unit (the front door, via the Defence Enterprise Committee) + Irish Coast Guard (IRCG)
Nexus Synergy Engagement Package — B01 · Irish Naval Service / Defence Forces RTI Unit (the front door, via the Defence Enterprise Committee) + Irish Coast Guard (IRCG)
STANDING BANNER — read before using this package. (1) Status honesty. Nexus Synergy Ltd is pre-incorporation, pre-revenue, pre-pilot. The Irish Naval Service, the Defence Forces RTI Unit, the Department of Defence and the Irish Coast Guard are TARGETS, not customers, partners, or deployments. MAHI is a prospective edge partner, not a signed one. Our ~178-surface UI gallery is a gallery, not traction. Nothing here is a contract, an LOI, or a commitment, and Patrick's "conditional cheque on a first LOI" is a genuine-but-conditional signal, not money in hand. (2) Anti-cookie-cutter. The §8 mockups are generated by
_build/archetypes/gen_naval-service-mahi-mda.pyand filled with this buyer's real entities (Irish EEZ ~450,000 km², a dark-vessel candidate near the CELTIC-1 cable corridor, the MAHI Alpha-8 USV edge, the C295 / LÉ tip-and-cue, the IRCG-117 / RNLI SAR picture). They are this buyer's screens, not a reused shell. (3) Every probability is[PRIOR]— a subjective pre-pilot estimate for prioritisation, never a forecast. Re-score after every real conversation. This is civil-security / maritime-domain-awareness framing, NOT warfighting; neutrality-compatible by construction.
§0 — Header + one-line thesis
Irish EEZ Maritime-Domain-Awareness COP — MAHI edge + Nexus Synergy sovereign brain. A sovereign, human-gated recognised maritime picture (RMP) that fuses sparse Irish sensors — AIS, Sat-AIS, the new €60m Thales sonar, C295 maritime-patrol returns, MAHI uncrewed-surface-vessel edge feeds and (optionally) the Sea-Scan fibre-sensing pilot — into one continuous, auditable picture of a 450,000 km² ocean the State physically cannot patrol, closing the loop detect → enrich → triage → human-gate → task → assess → close without ever resolving a person's identity.
One line: "You've just written in your own strategy that you need a national maritime security centre and a shared maritime picture you don't yet have, and you're buying the sensors that feed it. We are the EU-sovereign, Irish-built decision layer that fuses those sensors — above Thales and Airbus, not replacing them — with MAHI's uncrewed edge extending your reach without adding crew. Information instead of hulls."
§1 — Entity snapshot (cited)
| Attribute | Value | Source |
|---|---|---|
| Principal buyer | Irish Naval Service (under the Defence Forces / Department of Defence) — surveillance, EEZ patrol, fisheries protection, drug interdiction, SAR | gov.ie NMSS 2026-2030 [verified] |
| Realistic front door | Defence Forces RTI Unit / DRTIO (Defence Research, Technology & Innovation Office), reached via the Defence Enterprise Committee (DEC) — the first engagement surface for a pre-pilot startup | discovery-playbook §8 [verified route] |
| Civil wedge (faster) | Irish Coast Guard (IRCG) under the Dept of Transport — civil SAR + pollution + maritime-safety coordination; separate, lower-classification budget line | source card [verified] |
| Mission / scale | Protect Ireland's sovereign maritime interests across a ~450,000 km² EEZ (~10× the land area), one of the largest in the EU relative to fleet size | Irish Times / gov.ie [verified] |
| The written mandate | National Maritime Security Strategy (NMSS) 2026-2030, published Feb 2026: 41 objectives, six strategic objectives incl. protection of critical undersea infrastructure, and the concept of a "national maritime security centre" delivering a whole-of-government operational picture | gov.ie; RTÉ 25 Feb 2026; Naval Today 26 Feb 2026 [verified] |
| Sensors inbound (no brain) | €60m Thales towed-array sonar (3 OPVs, ~2027); Airbus C295 maritime-patrol aircraft + sonobuoys; Sea-Scan DAS+SoP fibre dark-vessel sensor (won the MSA challenge Feb 2026); €670m Bristow IRCG SAR aviation (live Aug 2025) | Irish Times 15 Jun 2025; TCD Feb 2026; RTÉ Aug 2025 [verified] |
| Prospective edge partner | MAHI (mahi.be, Belgium) — onboard autonomy + situational awareness for uncrewed surface vessels (MAHI Sense, Alpha-8 USV); defence references with UK MoD DSTL + US Navy TF59 | mahi.be; contact-register §B [verified person; terms unverified] |
| Status to us | Warm-leaning target: Don's direct MAHI founder relationship + Patrick Walsh's Defence-Forces network (RTI Unit via DEC); no NS/RTI named contact confirmed yet | source card [unverified — no contact made] |
§2 — The pain (web-verified, dated, cited)
- A tiny, shrinking fleet against an enormous ocean. Naval Service patrol days roughly halved 2020→2024 (to ~428 days); over half the fleet has been tied up at points; the service is >300 personnel short of an establishment of ~1,094. In Jan 2025 LÉ George Bernard Shaw deployed with its 76mm main gun non-functional — one ordnance technician left. The State cannot put enough hulls on 450,000 km², so it must substitute information for presence.
[verified — Irish Times 11 Jan 2025; Irish Examiner] - Sub-sea cable / seabed-infrastructure threat is now a named national-security concern. ~75% of transatlantic cables transit or pass near Irish waters. In 2025: a Russian-linked cargo ship dropped anchor near a cable in Irish waters; the intelligence ship Viktor Leonov was shadowed in the EEZ for ~36h; ~245 shadow-fleet vessels transited the EEZ 450+ times in seven months. Government responses: the €60m sonar buy, a cable-sabotage research programme (Dec 2025), and Ireland joining a naval-intelligence-sharing network (Apr 2025).
[verified — Irish Times; CSIS Ireland subsea case study; RTÉ Brainstorm 30 Jul 2025; Irish Times 5 Dec 2025] - No sovereign fusion layer today — the picture is to-be-built. The NMSS itself frames the "national maritime security centre" as a capability to establish; the Defence Forces is "working to establish a recognised maritime picture" — i.e. it does not yet have one. Sensors are being bought piecemeal (sonar, C295, Sea-Scan) with no published decision/COP layer to fuse them.
[verified — gov.ie NMSS; RTÉ 28 Feb 2026] - AI gap-filling is now official policy. The Defence Forces is explicitly using AI to cover personnel/resource shortfalls (incl. Sea-Scan AI vessel detection) — the institutional appetite for software force-multipliers is open.
[verified — Irish Times 16 May 2026] - The winner of the maritime-awareness grant is a sensor, not a picture. Sea-Scan (Trinity-ADAPT, led by Prof Marco Ruffini + Dr John Kennedy with DF Liaison Cdr Cathal Power) won the MSA challenge Feb 2026 using DAS + State-of-Polarisation fibre to detect dark vessels. It is a feed, not a fusion/decision layer — which leaves the COP slot open.
[verified — TCD Feb 2026; Engineers Ireland]
Net: Ireland has (a) written down that it needs a shared maritime picture it doesn't have, (b) started buying the sensors that will feed it, (c) endorsed AI as the way to cover its capacity gap, and (d) funded a sensor in the awareness grant — without owning the sovereign fusion/decision layer. That is the exact shaped hole this package fills.
§3 — Use-case & value (DECLINE-SAFE framing)
The bundle. MAHI provides the EU-sovereign edge (vessel/sensor autonomy, perception, fusion at the vessel — incl. uncrewed surface platforms that extend reach without adding crew). Nexus Synergy provides the EU-sovereign brain: a multi-vessel / multi-domain COP that fuses MAHI edge feeds + the new sonar + C295/sonobuoy + AIS + Sat-AIS + satellite + (optionally) Sea-Scan into one recognised maritime picture, with the closed action loop on top — over a customer-editable maritime ontology.
Value propositions:
- Substitute information for hulls. Anomaly detection on the EEZ (dark vessels, AIS spoofing/loitering near cable corridors, shadow-fleet pattern-of-life) so the few available assets are vectored only where it matters.
- Sub-sea cable watch. A cable-corridor geofence that correlates anchor-drop / loiter behaviour near the cable landings + the €60m sonar returns + Sea-Scan DAS, raising one triaged alert with a human-gated tasking recommendation. Directly serves the NMSS undersea-infrastructure objective and the Dec-2025 cable-research programme.
- The "national maritime security centre," in software. Civil (IRCG, SFPA, EMSA/CISE) + military (Naval Service, Air Corps) feeds in one shared, classification-partitioned picture — exactly the whole-of-government centre the NMSS proposes. The classification-aware AI Provider Router + conformity-by-construction (FRIA + provenance ledger + transparency portal) make it defensibly sovereign and audit-ready.
- SAR coordination for IRCG. A multi-asset SAR common picture + tasking write-back for the new Bristow aviation fleet and the RNLI — a lower-classification, faster-to-pilot civil wedge (the second mockup, §8).
DECLINE-SAFE — explicitly out of scope (and contractually excluded):
- NOT predictive policing · NOT real-time/biometric identification · NOT emotion recognition · NOT social scoring · NOT untargeted scraping · NOT psychometric profiling.
- This is maritime-domain-awareness / coast-guard / critical-infrastructure protection — dual-use civil security, not warfighting. The COP works on vessel tracks and infrastructure telemetry, not persons. Vessel identity is pseudonymised in the UI; crew-identity resolution is structurally forbidden without a Tier-4 warrant (enforced as a DB
CHECK, §5). The escalation ceiling is a Naval/Coast-Guard vector or VHF challenge under maritime-safety authority, never SIGINT collection on individuals. - Neutrality-compatible: sold as situational awareness and infrastructure protection, never force projection — which fits Irish military-neutrality sensibilities and keeps the whole engagement clear of every Declined-List item.
§4 — Ontology (this buyer's domain entities + relationships)
Entities
Vessel (per MMSI + IMO + flag-state + owner-chain where known)
AIS-Contact (per-ping LatLon + heading + SOG + spoof-suspicion + receive-station)
SatAIS-Contact (Copernicus/EMSA pass-geometry confidence)
SAR-Scene-Match (Sentinel-1 IW; centroid LatLon; backscatter; estimated LOA)
USV-Edge-Cue (MAHI Alpha-8: EO/IR perception cue, on-scene, no crew aboard)
SubseaCable-Seg (CELTIC-1 corridor; criticality weight; landfall)
Sonar-Contact (Thales towed-array, OPV, ~2027; bearing + classification)
SAR-Distress (IRCG: distress LatLon, casualty count, drift cone)
Rescue-Asset (IRCG-117 helo, RNLI ALB, C295, LÉ OPV; posture + ETA + fit)
Sanctions-Entry (CFSP framework; per-vessel / per-flag pattern)
Relationships
Vessel --ais_telemetry--> AIS-Contact
AIS-Contact --contradicted_by--> {SatAIS-Contact, SAR-Scene-Match, Sonar-Contact}
USV-Edge-Cue --corroborates--> AIS-Contact (MAHI edge visual)
AIS-Contact --loiters_near--> SubseaCable-Seg
Vessel --sanctioned_via--> Sanctions-Entry (where present)
SAR-Distress --assigned--> Rescue-Asset (Hungarian optimum, human-gated)
The ontology is customer-editable: the Naval Service MOC can add a HF-Radar-Contact entity or re-weight SubseaCable-Seg criticality without a vendor change-request — the differentiator versus a prime's fixed sensor-bolt-on dashboard.
§5 — Data model (2 synergy.* tables, RLS + a load-bearing CHECK)
Entities are org_id-scoped for the app.current_org_id RLS predicate (Naval Service MOC tenant ≠ IRCG tenant ≠ a Member-State Navy tenant). The load-bearing safety property is that no crew-identity / SIGINT escalation may be written without a Tier-4 warrant reference — the civil-security boundary, enforced in the database, not just the UI.
SQL43 lines-- Dark-vessel candidate from AIS-vs-SatAIS-vs-SAR contradiction + MAHI USV edge cue, org-scoped + RLS. CREATE TABLE synergy.dark_vessel_candidate ( id uuid PRIMARY KEY DEFAULT gen_random_uuid(), org_id uuid NOT NULL, -- RLS: app.current_org_id mmsi text NOT NULL, -- '308194632' flag_state text, -- 'CMR' self_report_pos geography(POINT, 4326) NOT NULL, -- 51.4N 11.2W (broadcast AIS) satais_pos geography(POINT, 4326), -- 51.0N 11.9W (Copernicus EMSA) delta_nm numeric(6,2) NOT NULL DEFAULT 0, -- 41.00 spoof_score numeric(4,3) NOT NULL, -- 0.910 sar_link_weight numeric(6,2), -- Fellegi-Sunter W = 9.62 usv_edge_cue boolean NOT NULL DEFAULT false, -- MAHI Alpha-8 EO/IR corroboration near_cable_segment text, -- 'CELTIC-1' obs_level text NOT NULL DEFAULT 'inferred' CHECK (obs_level IN ('asserted','reported','inferred','direct','confirmed')), crew_id_warrant_ref text, -- Tier-4 warrant; NULL until issued sigint_escalated_at timestamptz, -- NULL while Tier-3 civil COP prov_o jsonb NOT NULL, -- LOAD-BEARING INVARIANT: no crew-identity / SIGINT escalation without a Tier-4 warrant. CHECK (sigint_escalated_at IS NULL OR crew_id_warrant_ref IS NOT NULL) ); ALTER TABLE synergy.dark_vessel_candidate ENABLE ROW LEVEL SECURITY; CREATE POLICY org_isolation ON synergy.dark_vessel_candidate USING (org_id = current_setting('app.current_org_id')::uuid); -- Human-gated tasking order (Naval vector / IRCG SAR / VHF challenge); execute only after dual-control sign-off. CREATE TABLE synergy.tasking_order ( id uuid PRIMARY KEY DEFAULT gen_random_uuid(), org_id uuid NOT NULL, candidate_id uuid REFERENCES synergy.dark_vessel_candidate(id), task_kind text NOT NULL CHECK (task_kind IN ('visual_identify','vhf_challenge','sar_vector','sonar_cue')), asset_id text NOT NULL, -- 'CASA-235-CM252' | 'IRCG-117' | 'RNLI-ARAN-ALB' lawful_basis text NOT NULL, -- 'Maritime Safety Act 2005' | 'Defence Act 1954 s6' sign1_actor text, sign1_at timestamptz, -- MIO duty sign2_actor text, sign2_at timestamptz, -- N2 OOW / IRCG duty coordinator executed_at timestamptz, -- NULL until BOTH signs present prov_o jsonb NOT NULL, -- LOAD-BEARING INVARIANT: an order cannot execute without 2-of-2 human dual-control sign-off. CHECK (executed_at IS NULL OR (sign1_at IS NOT NULL AND sign2_at IS NOT NULL)) ); ALTER TABLE synergy.tasking_order ENABLE ROW LEVEL SECURITY; CREATE POLICY org_isolation ON synergy.tasking_order USING (org_id = current_setting('app.current_org_id')::uuid);
The second CHECK is the one that makes the demo land with a sovereign buyer: the database itself refuses to record an executed tasking order unless two named humans signed it. Restraint is structural, not a policy promise.
§6 — Action-loop pseudocode (detect → enrich → triage → gate → task → execute → BDA → close)
Python42 linesdef assess_eez_anomaly(candidate, requesting_org): # 1. DETECT — AIS-vs-SatAIS delta + spoof score (workflow-DAG-composable spoofing-detection Skill). if candidate.spoof_score < 0.7: return Disposition.WATCH # show on globe, no tasking # 2. ENRICH — SAR-scene Fellegi-Sunter link + MAHI USV edge EO/IR cue. candidate.sar_link_weight = fellegi_sunter_sar_to_track(candidate) # W = 9.62 (auto-match > T_mu=4.0) candidate.usv_edge_cue = mahi_alpha8.request_visual(candidate) # edge corroboration, no crew if candidate.sar_link_weight is None and not candidate.usv_edge_cue: return Disposition.WATCH # 3. TRIAGE — Tier-3 proportionality; 5-persona Decision Room incl. a civil-society reviewer. P = proportionality(necessity=0.91, proport=0.94, lawful=0.97) # P = mean(N,P,L) = 0.84 gate = decision_room( proposal = TipAndCue(candidate, asset="CASA-235-CM252", task="visual_identify", rules="no crew-ID"), personas = [OP_TASKING, PROPORTIONALITY, CIVIL_SOCIETY('Foxglove'), DEVILS_ADVOCATE, SATAIS_ANALYST], threshold = P_MIN_TIER3) # 4. GATE — the civil-security boundary is enforced here AND in the DB CHECK. if gate.requests_crew_id and candidate.crew_id_warrant_ref is None: return Disposition.DEFER_TO_TIER4 # CHECK forbids the write; warrant seat instantiates if not gate.passed: return Disposition.REFUSED(gate.rationale) # logged with full transcript # 5. TASK + 6. EXECUTE — only after 2-of-2 dual-control sign-off (DB CHECK). order = TaskingOrder(candidate, task="visual_identify", asset="CASA-235-CM252", lawful_basis="Maritime Safety Act 2005") require_dual_control(order, signers=["MIO-duty", "N2-OOW"]) # 2-of-2 or no execute write_back("naval_cad", requesting_org, order) # CoT 2.0 mission-tasking write_back("eurosur.alert", "atlantic_approaches", EurosurAlert(candidate, obs_level="inferred")) # escalates on BDA # 7. BDA — C295 / MAHI USV on-station raises ObservationLevel. bda = await_bda(order) # visual confirm -> 'confirmed' candidate.obs_level = bda.observation_level if bda.confirmed: register_prospective(candidate, trigger="shadow_fleet_AIS_spoof_pattern_Atlantic", alpha=0.4) # 8. CLOSE — Merkle-anchored audit envelope; sanctions lodge stays STAGED behind 'confirmed'. audit.append(candidate.id, boundary_state="tier3-civil-COP", merkle=True) return Disposition.EXECUTED
§7 — nexus-workflows YAML DAG (the same loop, declarative, with a blocking human-gate node)
YAML50 lines# nexus-workflows DAG — eez-mda-anomaly-loop (NO cron anywhere; this IS the scheduler) workflow: eez-mda-anomaly-loop tenant: naval-service-moc trigger: on_event: ais.anomaly.candidate # emitted by the spoofing-detection Skill nodes: - id: detect skill: ais-spoof-classifier out: { spoof_score: $.score } gate: { expr: "spoof_score >= 0.7", else: watch } - id: enrich_sar skill: fellegi-sunter-sar-to-track out: { sar_link_weight: $.W } # 9.62 -> auto-match - id: enrich_usv skill: mahi-alpha8-visual-cue # MAHI edge, no crew aboard out: { usv_edge_cue: $.cue } - id: triage_decision_room skill: decision-room-compose personas: [op_tasking, proportionality, civil_society_foxglove, devils_advocate, satais_analyst] out: { proportionality: \$.P, reco: \$.reco } - id: human_gate # <<< BLOCKING human-gate node >>> type: manual_approval blocking: true require: dual_control # 2-of-2 approvers: [mio_duty, n2_oow] forbid_without: tier4_warrant # crew-ID escalation is refused here on_reject: { goto: close_refused } - id: task_writeback after: human_gate skill: cad-writeback # Naval CAD (CoT 2.0) + eurosur.alert post lawful_basis: "Maritime Safety Act 2005 ; Defence Act 1954 s6" - id: bda after: task_writeback skill: bda-capture # C295 / USV on-station -> obs_level=confirmed out: { obs_level: $.level } - id: close after: bda skill: case-close audit: { merkle: true, anchor: sigstore-rekor } prospective_item: { trigger: shadow_fleet_AIS_spoof_pattern_Atlantic, alpha: 0.4 } - id: close_refused skill: case-close audit: { merkle: true, note: "refused at human_gate ; transcript retained" }
The human_gate node is blocking: true with require: dual_control — the DAG cannot proceed to task_writeback without two named approvals, and forbid_without: tier4_warrant mirrors the DB CHECK so the safety property is enforced in two independent places.
§8 — UI/UX mockups (VERBATIM generated ASCII)
Figure 8.1 — globe_rmp (primary, CesiumJS-globe). The Naval Service MOC seat: ShellLayout + TopBar with PccPill (rightSlot) + RendererChip reading CesiumJS-globe + LeftSidebar + MapConsole rendering the top-down Irish-EEZ overview (EMODnet bathymetry, the 200nm boundary arc, the ASTRA-MARIA-7 self-report icon, the 41nm delta to the Sat-AIS position, the Sentinel-1 SAR footprint, the CELTIC-1 cable corridor, the MAHI Alpha-8 USV edge marker) + a right-docked ChatTerminal hosting the TRACKS rail and the five-persona Decision Room + HistoryRail + BottomStatusBar + ClassificationBanner. SVG twin: _build/figures/naval-service-mahi-mda/uc-globe_rmp.svg.

Figure — Recognised maritime picture (globe_rmp). Production-fidelity React surface (buildable); the faithful ASCII follows.
+--------------------------------------------------------------------------------------------------+
| Irish EEZ Recognised Maritime Picture - Naval Service MOC + MAHI edge (CesiumJS-globe) |
+-------------------------------------------------------------------+------------------------------+
| . - '' - . IRISH EEZ ~450,000 km2 (2nd-largest in EU) | TRACKS (sorted: spoof) |
| / o ----._ \ self-report 51.4N 11.2W (broadcast AIS) | ASTRA-MARIA-7 CMR flag |
| | [!] AIS spoof | EMODnet bathy -1840m :: IHO S-101 ENC | Sat-AIS vs self: 41nm |
| | Sat-AIS 51.0N | M/V ASTRA-MARIA-7 (CMR flag, AIS gaps) | SAR-confirm W=9.62>Tmu |
| | 41nm SW delta | -> dark-vessel candidate (spoof 0.91) | loiter 6nm N CELTIC-1 |
| \ S-1 SAR match / Sentinel-1 IW @14:32Z confirms 51.0N 11.9W | MAHI Alpha-8 USV (edge) |
| \ '. CELTIC-1 .'/ loiters 6nm N of CELTIC-1 cable corridor | EO/IR cue, no crew aboard |
| ' - x__ - ' MAHI Alpha-8 USV edge: EO/IR visual cue | 5 more dark-cands T3 |
| Camera {53.20N -14.50W alt 850km pitch -90} CesiumJS-globe | 2,140 AIS contacts live |
| Concurrence: 41nm AIS-vs-SatAIS + SAR W=9.62 + cable-loiter | [#] Sentinel-1 IW @14:32 |
| -> ANOMALY 0.91 :: HUMAN GATE before any Naval/IRCG tasking | [#] C295 tip-and-cue 4h |
| | [#] LE-WB-SHAW standby |
| | gate: vector + VHF query |
| | NO crew-ID (civil COP) |
+-------------------------------------------------------------------+------------------------------+
| EMODnet bathy + IHO S-101 + AIS(terr+SatAIS) + MAHI USV edge + Thales sonar :: NMSS centre-in-SW |
+--------------------------------------------------------------------------------------------------+
Figure 8.2 — map_ops (secondary, IRCG SAR coordination, MapConsole 2D). The IRCG virtual-JRCC seat: ShellLayout + TopBar/RendererChip (2D MapLibre) + a LeftSidebar LAYERS panel (with a gated Crew identity (Tier-4 warrant) toggle that is off-by-construction) + MapConsole showing the distress, the Met-Éireann drift cone, the IRCG-117 / RNLI / MAHI-USV assets, the HLS + ED-capacity overlay, and the golden-hour clock + Inspector + BottomStatusBar incident ticker.

Figure — Operational picture (map_ops · live MapLibre). Production-fidelity React surface (buildable); the faithful ASCII follows.
+------------------------------------------------------------------------------------------------------+
| IRCG SAR Coordination - virtual-JRCC operating picture (MapConsole 2D) |
+--------------------+---------------------------------------------------------------------------------+
| LAYERS | MAP ========================================================================== |
| [x] AIS (terr+sat) | Galway Bay -> SW approaches :: IRCG virtual-JRCC SAR slice |
| [x] MAHI USV edge | .--drift cone (Met Eireann ROMS, SSW 1.2kt)--. |
| [x] EMODnet bathy | / [D] M/V CELTIC-VENTURE distress 53.04N 9.66W \ |
| [x] S-101 ENC | | IRCG-117 helo (Shannon) ETA 38m hoist-capable | |
| [x] Cable corridors| | RNLI Aran ALB x2 ETA 22m / 41m | |
| [x] Drift model | | MAHI Alpha-8 USV on-scene EO/IR relay, no crew | |
| [x] RNLI stations | \ HLS: Galway Airport :: ED capacity: UHG 4 beds / |
| [x] IRCG stations | '----- 6nm grid ----- golden-hour clock 00:47 -----' |
| [ ] Thales sonar | Camera {lat 53.04 lon -9.66 alt 4.2km pitch -25} |
| [ ] Sentinel-1 SAR | Tasking write-back: IRCG MRCC + RNLI pager (human-gated) |
| [x] SAR assets | |
| --- gated --- | |
| [ ] Crew identity | |
| (Tier-4 warrant) | |
+--------------------+---------------------------------------------------------------------------------+
| INCIDENTS: CELTIC-VENTURE distress 53.04N 9.66W | IRCG-117 38m | RNLI x2 | MAHI USV on-scene | gate |
+------------------------------------------------------------------------------------------------------+
§8b — Field-unit (Pixel) surfaces
The same scenario on the Pixel 10 Pro Fold field unit (Nexus Field app), tightly coordinated with the dashboard COP above — command pushes the task, the unit accepts + ACKs, shares position and reports back to the COP. Built on the same synergy.field_unit / field_task / field_report contract; see §9 and the cluster coordination composite.
Figure §8b.1 — Folded cover · tasking glance (ground_glance): the incoming IMMEDIATE task, ACCEPT + ACK, alert chips, bearing-to-objective.
Figure §8b.2 — Unfolded inner display · field COP: two-pane mini-map + task list + teammate roster + air/command coordination + PTT, with the Material-3 NavigationBar + Report FAB.
§9 — UI/UX flow (the literal click-path)
Click-path (globe_rmp → gate → task):
MapConsole dark-vessel diamond (red) → click opens Inspector/TRACKS for ASTRA-MARIA-7 → press [Decision Room] → five persona bubbles populate in ChatTerminal (the Devil's-advocate "coverage gap" objection downgrades the reco; the Sat-AIS analyst restores it with 96% detect-confidence + the Sentinel-1 fix) → operator reads the civil-society red line (no crew-ID) → [/dual-control sign — MIO] ✓ then [/dual-control sign — N2 OOW] ✓ → [/execute → C295 tip-and-cue] → BdaCaptureBubble raises ObservationLevel on visual confirm → [Close + register ProspectiveItem].
[Globe RMP] --click diamond--> [TRACKS/Inspector] --[Decision Room]--> [5 persona bubbles]
| |
| reco DOWN (devil) -> UP (analyst)
v v
[civil-society red line: NO crew-ID] --2-of-2 sign--> [HUMAN GATE] --pass--> [/execute C295]
| |
fail -> [Refuse + log] [BDA visual confirm]
|
[Close + ProspectiveItem]
For the IRCG SAR slice (map_ops): distress marker → [Assign assets] runs the Hungarian optimum (IRCG-117 + RNLI ×2 + MAHI USV) → duty coordinator dual-control sign → [/execute → IRCG MRCC + RNLI pager] → golden-hour clock tracks to BDA.
§10 — Decision-Room transcript (the gated decision)
◀◀ EU-CONFIDENTIAL · TLP:AMBER · T3 ▶▶ [Nexus-Synergy rc10 · tactical] MapConsole=CesiumJS-globe
┌──────────────────────────────────────────────────────────┬──────────────────────────────────────┐
│ MapConsole — Irish EEZ ~450,000 km² │ Nexus Chat — Decision Room [▢][▭][✕]│
│ ▣ 2,140 AIS contacts live · 6 dark candidates ◆T3 │ ⊟ Op-tasking [L3 Mistral-Large] │
│ ▣ ASTRA-MARIA-7 MMSI 308194632 (CMR) │ Reco: C295 tip-and-cue next pass + │
│ Sat-AIS contradicts self-report by 41nm │ visual-identify 4h; register shadow-│
│ ▣ Sentinel-1 IW @14:32:18Z confirms a hull 51.0N 11.9W │ fleet ProspectiveItem. NO crew-ID. │
│ ▣ MAHI Alpha-8 USV on-scene: EO/IR cue, no crew aboard │ ⊟ Proportionality [L3] P=0.84 │
│ ▣ loiters 6nm N of CELTIC-1 cable corridor │ N0.91 P0.94 L0.97 → mean 0.84. │
│ │ Tier-3 vessel-tracking proportionate│
│ │ NO individual targeted. [Direct] │
│ │ ⊟ Civil-society (Foxglove flavour) │
│ │ Sat-AIS at Tier-3 is fine. RED LINE:│
│ │ do NOT resolve crew identity / link │
│ │ to persons without a Tier-4 warrant │
│ │ if the vessel makes an Irish port │
│ │ call. Vessel telemetry ≠ people. │
│ │ ⊟ Devil's advocate [L3] │
│ │ A 41nm AIS-vs-SatAIS delta is ALSO │
│ │ consistent with a Sat-AIS coverage │
│ │ gap. Do not advance shadow-fleet │
│ │ attribution on a gap. │
│ │ ⮑ RECO DOWNGRADED → "monitor only" │
│ │ ⊟ Sat-AIS analyst [L3 EuroLLM] │
│ │ Gap REJECTED: Sat-AIS-3 pass @14:30Z│
│ │ covered this cell at 96% detect-conf│
│ │ + Sentinel-1 IW puts a hull at │
│ │ 51.0N 11.9W ±0.4nm + MAHI USV EO/IR │
│ │ cue. THREE sources agree. [Direct] │
│ │ ⮑ RECO RESTORED → C295 APPROVED │
│ │ ─────────────────────────────────── │
│ │ [/dual-control sign — MIO duty] ✓ │
│ │ [/dual-control sign — N2 OOW] ✓ │
│ │ GATE: ACCEPT P=0.84 ≥ P_min · 2-of-2│
│ │ [/execute → C295 visual-identify] │
│ │ [/execute → EUROSUR alert post] │
├──────────────────────────────────────────────────────────┴──────────────────────────────────────┤
│ HistoryRail: [● 14:32 6 dark-cands][ASTRA spoof 0.91][reco ↓ devil / ↑ analyst][MAHI USV cue] │
│ BottomStatusBar: sync✓ · NORMAL · Galileo-OS · OneWeb (sovereign) · L3 Mistral-Large │
◀◀ EU-CONFIDENTIAL · TLP:AMBER · T3 ▶▶
The dissent did real work. The Devil's-advocate persona downgraded the reco to "monitor only" on a coverage-gap steel-man; the Sat-AIS analyst rejected the gap with three corroborants (96% detect-confidence, the Sentinel-1 IW fix, and the MAHI USV EO/IR cue) and restored the C295 tasking. The civil-society (Foxglove-flavour) reviewer holds a hard red line — no crew-identity resolution without a Tier-4 warrant — which the DB CHECK enforces so it cannot be overridden by an operator click. The gate verdict (ACCEPT, P=0.84, 2-of-2) and the ↓↑ swing are logged in HistoryRail for replay.
§11 — Write-back + BDA + PROV-O chain + deltas-only regulatory traceback
Write-back (three targets, each with named lawful basis + dual-control gate):
- Naval Service CAD — C295/CASA-235 tasking via CoT 2.0 mission-tasking extension; lawful basis = Defence Act 1954 §6 + Maritime Safety Act 2005.
eurosur.alert— European Border Surveillance alert post in CISE message format; lawful basis = Regulation (EU) 1052/2013 (EUROSUR). ObservationLevelinferred→ escalates on BDA.sanctions.cfsp.lodge— STAGED behind BDA-confirm; lawful basis = SI implementing CFSP measures; bilateral with the DFA Sanctions Unit. It does not fire on the inferred attribution.
BDA (loop close). C295 on-station + MAHI USV EO/IR confirm a vessel matching the SAR-estimated dimensions + a flag-of-convenience pattern; the BdaCaptureBubble raises ObservationLevel inferred → confirmed, which unlocks the staged sanctions lodgement. ProspectiveItem registered: shadow-fleet-AIS-spoof-pattern-Atlantic (α=0.4; cross-tenant opt-in recorded as a PROV-O data-sharing edge).
PROV-O attribution chain (Turtle-style):
Plain Text22 lines:AisTrack-308194632 a prov:Entity ; prov:wasGeneratedBy :AisMultiplexerIngest . :SatAisCross a prov:Entity ; prov:wasGeneratedBy :SatAisCorrelator ; prov:used :CopernicusEmsaFeed . :Sentinel1Match a prov:Entity ; prov:wasGeneratedBy :SarSceneMatch ; prov:used :Sentinel1IwScene-143218Z . :UsvEdgeCue a prov:Entity ; prov:wasGeneratedBy :MahiAlpha8Perception . # edge, no crew :SpoofScore-0.91 a prov:Entity ; prov:wasGeneratedBy :AisSpoofClassifier-v1.2 ; prov:used :AisTrack-308194632, :SatAisCross, :Sentinel1Match . :SarLink-W9.62 a prov:Entity ; prov:wasGeneratedBy :FellegiSunterLinker ; prov:used :Sentinel1Match, :AisTrack-308194632 . :DecisionRoom-B01 a prov:Entity ; prov:wasGeneratedBy :DecisionRoomCompose ; prov:used :SpoofScore-0.91, :SarLink-W9.62, :UsvEdgeCue ; prov:wasAttributedTo :OpTasking, :Proportionality, :CivilSociety, :DevilsAdvocate, :SatAisAnalyst . :RecoSwing-down-up a prov:Entity ; prov:wasDerivedFrom :DevilsAdvocate, :SatAisAnalyst . :DualControl-B01 a prov:Entity ; prov:wasGeneratedBy :DualControlGate ; prov:wasAttributedTo :MIO-duty, :N2-OOW . :C295Tasking a prov:Entity ; prov:wasGeneratedBy :CADWriteBack ; prov:wasDerivedFrom :DecisionRoom-B01, :DualControl-B01 . :BDA-B01 a prov:Entity ; prov:wasGeneratedBy :C295VisualConfirm ; prov:used :C295Tasking . :SanctionsLodge-B01 a prov:Entity ; prov:wasGeneratedBy :CfspLodgeAdapter ; prov:wasDerivedFrom :BDA-B01 . # staged behind 'confirmed' :CaseClose-B01 a prov:Entity ; prov:wasGeneratedBy :CaseCloseActivity ; prov:used :BDA-B01, :SanctionsLodge-B01 .
Chain unbroken: finding → enrich (Sat-AIS + SAR + MAHI USV) → score → ER-link → 5-persona triage (+ reco-swing) → dual-control signers → execute (CAD + EUROSUR) → BDA → staged sanctions lodge → close. Every node is replayable.
Regulatory traceback (deltas only — what differs from the §4.0 defaults):
- AI Act Arts 6 + 27 (high-risk + FRIA) engaged because the use is law-enforcement-adjacent maritime surveillance; Art 5 prohibited practices (applicable 2 February 2025) are not triggered — no biometric/predictive/social-scoring path exists.
- LED (Dir (EU) 2016/680) Art 8 — N/A: the Naval Service is lead (not An Garda Síochána) and no identified-crew personal data is processed; the Foxglove red line + the DB
CHECKkeep it N/A. - EUROSUR (Reg (EU) 1052/2013) governs the alert-post write-back; CFSP (Council Decision 2014/512/CFSP + successors) governs the staged sanctions lodge.
- NIS2 (Dir (EU) 2022/2555) engaged on the defensive side because the contact loiters near a designated cable corridor — critical-infrastructure protection, never offensive action.
§12 — Buyer & stakeholders
| Role | Who / where | Notes |
|---|---|---|
| Capability / budget owner | Department of Defence + Defence Forces (Naval Service for the sea-going piece) | NMSS 2026-2030 is the written mandate; DoD owns the "national maritime security centre" concept. [verified] |
| R&D / innovation front door | Defence Forces RTI Unit / DRTIO, reached via the Defence Enterprise Committee (DEC) | The realistic first engagement surface for a pre-pilot startup. [verified route] |
| Named MDA-innovation insider | Cdr Cathal Power — DF Liaison on the Sea-Scan MSA project | A known, named maritime-innovation insider to map (not yet a contact — do not invent a relationship). [verified name] |
| Civil SAR / maritime safety | Irish Coast Guard (Dept of Transport) | Separate, faster, lower-classification pilot path. [verified] |
| Edge partner (prospective) | Pieter-Jan Note (CEO, MAHI) — founder-to-founder; Bertold Van den Bergh / Quinten Lauwers (tech integration) | The gate-resolver: MAHI as contracting prime lifts V 0.10→0.75. [verified persons; terms unverified] |
| Warm-intro sponsor | Patrick Walsh (CEO, Dogpatch Labs) | Defence-Forces network + conditional cheque on a first LOI. [verified — patrick@dogpatchlabs.com] |
| Buyer champion | [Named contact TBD] — the NS/RTI owner of the centre pilot budget | Patrick + Don's MAHI relationship to identify; do not assert before confirmed. |
| EU plumbing the COP must speak to | EMSA / CISE, EUROSUR, EFCA, REPMUS | Integration surfaces, not buyers (this round). [verified] |
§13 — Competition / incumbency + comparator
- No entrenched sovereign-COP incumbent in Ireland. The recognised maritime picture is explicitly to-be-built — there is no Palantir-Maritime / big-SI lock-in to displace today. Favourable.
[verified — gov.ie NMSS; RTÉ 28 Feb 2026] - Sensor + SI gravity is the real risk. Thales (won the €60m sonar) and Airbus (C295) are deep in the Irish maritime stack and could push their own fusion/C2 software downward; an SI could bundle a fusion layer into a hardware framework and lock us out.
[verified — Irish Times 24 Jan 2026 "one country and company gain the most"] - Sea-Scan / Trinity-ADAPT, MAHI, others = feeds, not foes. Position at the fusion altitude; a crowded sensor field increases the need for a neutral fusion layer.
- Comparator: Palantir MetaConstellation / Maven Smart System for MDA. Publicly documented satellite-tasking + fusion, but US-foundation-model in the loop and no published RMP-spoofing-detection-with-civil-society-dissent composition. Our wins: sovereign-EU LLMs throughout (Mistral-Large + EuroLLM, no US model), sovereign SatCom (OneWeb selected over Starlink by policy), a customer-editable ontology, a genuine closed action loop, and the civil-society red line enforced in the database. Honest loss: Sentinel-1 revisit (~2-12h) is a hard tip-and-cue floor no platform fixes, and ours doesn't either; sub-hour revisit needs commercial SAR tasking, which re-opens a sovereignty question we deliberately decline.
[verified comparator existence; capability claims AS-PUBLISHED]
§14 — Readiness (honest, pre-pilot)
- Reuse (real): the architecture a sovereign buyer needs to hear is genuinely built as a thesis — classification-aware AI Provider Router, FRIA + provenance ledger + transparency portal, the 5-persona Decision Room, the editable ontology, the triple-renderer map (CesiumJS-globe for EEZ scale). The W-12 RMP pattern (AIS-multiplexer + Sat-AIS cross-check) is AS-IMPLEMENTED at the ingest level.
- Real gaps:
- The MAHI data-contract spike is unbuilt — the MAHI Sense open-API / track schema / ROC bus needs a technical session and a scoped integration before any joint pitch is credible.
- Turning the gallery into a fielded, data-fed maritime COP (live MAHI/AIS/sonar/satellite feeds, classification partitioning, audit ledger) is real engineering work ahead.
- The workflow-DAG-composable spoofing-detection Skill is AS-DESIGNED, not fielded; the Fellegi-Sunter SAR-to-track linker is specified, not deployed.
- No NS/RTI named contact is confirmed; the centre pilot-budget owner is unknown.
- Honest scoping rule: a pilot must start narrow (one cable corridor watch, or the IRCG SAR-COP) — not "the whole national centre."
§15 — ENGAGEMENT PLAYBOOK (first-contact → free discovery → demo → pilot → LOI → contract)
Stage 0 — Verify (pre-engagement). Trigger: before any intro is requested. Owner: Don. Activity: confirm MAHI bundle terms (edge scope, open-API/data-contract reality, any Ultranav exclusivity) via Don↔Pieter-Jan; confirm Patrick's specific NS/RTI contact; confirm whether IRCG (Transport) is the faster first pilot. Exit: MAHI willing to act as contracting prime + a named NS/RTI or IRCG target. Deliverable: a one-line bundle-terms memo + a named target.
Stage 1 — First contact (warm intro). Trigger: Stage 0 done. Owner: Patrick (NS/RTI via DEC) + Don (MAHI founder-to-founder). Activity: send the two warm-intro emails below. Exit: a discovery call booked. Deliverable: calendar invite with the discovery agenda attached.
Warm-intro email — to Patrick (adapted from outreach-drafts §1): Subject: One named intro to turn the maritime-COP thesis into a first MOU Patrick — you said the LOI is the unlock, so here is the single named ask that matters most for the maritime play: an introduction to the Defence Forces RTI Unit (via the DEC) — ideally whoever is mapping the NMSS "national maritime security centre" — so I can show, on Ireland's own EEZ, a sovereign recognised-maritime-picture that fuses the sensors the State is already buying (the €60m sonar, the C295, Sea-Scan) into one human-gated, fully-audited picture. I'm walking in with MAHI (Belgian uncrewed-edge partner, DSTL/TF59 references), so this is an edge+brain team, not slideware. Free discovery, no cost to them, explicitly civil-security MDA — not warfighting. Happy to walk you through the demo first so you're comfortable putting your name to it.
Warm-intro email — to Pieter-Jan Note (MAHI) (adapted from outreach-drafts §2): Subject: Edge + sovereign brain — the Irish NMSS window Pieter-Jan — the Irish timing is unusually good: the first-ever NMSS (Feb 2026) commits to a national maritime security centre + undersea-cable protection, and the Naval Service can't crew its EEZ. Your MAHI Sense edge feeds detect/sense; we add enrich → triage → human-gate → task → assess → close across a fleet, writing tasking back through your open API. Two asks: (1) a technical session on the MAHI Sense API / track schema to scope the data contract; (2) an honest conversation about MAHI acting as the contracting vehicle for a first Irish discovery/pilot — candidly, that clears a legal gate on my side and we both win.
Stage 2 — Free discovery (60 min, no commitment). Trigger: call booked. Owner: Don (+ MAHI optional). Activity: run the discovery agenda — frame & permission (0:00-0:05) → current operating picture / surface the pain (0:05-0:20) → governance, lawful basis, sovereignty (0:20-0:35) → stakeholders, budget, timing (0:35-0:45) → map their loop onto our 8-stage action loop (0:45-0:55) → close & earn the demo (0:55-1:00). No slides, no money ask. Exit: named sponsor agrees there is a problem worth solving + asks for the demo. Deliverable: same-day discovery write-up (sponsor, top-3 pains in their words, lawful basis/classification, action-loop gap, Declined-List PASS, funding vehicle hinted, next step + date).
Stage 3 — Scoped demo (their loop, our platform). Trigger: discovery exit met. Owner: Nexus Synergy + MAHI. Activity: replay their EEZ anomaly + cable-corridor watch + human-gated tasking on the CesiumJS-globe RMP, with the MAHI edge feed mocked or live; lead with the sovereignty differentiators that matched their qualification answers (AI Provider Router, FRIA/provenance, the 5-persona Decision Room with the civil-society reviewer seated inside the decision, claim-level epistemic transparency, the published Declined List). Exit: "could we try this on a narrow, real slice?" Deliverable: the one-page pilot proposal (Stage 4).
Stage 4 — Pilot (time-boxed, grant-funded). Trigger: demo exit ask. Owner: Joint. Activity: run the narrow pilot (one cable corridor OR IRCG SAR-COP). Exit: success criteria met or credibly trending. Deliverable: the filled 1-page pilot proposal below.
1-page pilot proposal — filled for THIS buyer:
- Title & sponsor: Irish EEZ Cable-Corridor Watch — sovereign MDA fusion pilot. Sponsor: [NS/RTI or IRCG — TBD]; budget owner: [TBD].
- The problem (their words): "We have sensors arriving and a 450,000 km² EEZ we can't patrol; we don't have one shared picture, and we reconstruct cable incidents after the fact."
- The one thing we'll prove: We can fuse AIS + Sat-AIS + a Sentinel-1 SAR scene (+ a mocked/live MAHI USV cue) into one human-gated cable-corridor picture and cut time-to-detect-anomaly from X to Y. (X/Y = baseline TBD, measured in week 1.)
- Scope IN/OUT: IN — one cable corridor (e.g. CELTIC approaches), globe_rmp + Decision Room + dual-control gate. OUT — the whole national centre, any crew-identity resolution, any Declined-List capability (explicitly excluded).
- Data & lawful basis: AIS (open) + Sat-AIS (Copernicus/EMSA) + Sentinel-1 (open) + mocked MAHI feed; classification EU-RESTRICTED; EU-hosted; lawful basis Maritime Safety Act 2005 / Defence Act 1954 §6.
- Success criteria: 2-4 sponsor-agreed metrics (time-to-detect; false-alarm rate; auditability — every decision replayable). Baselines TBD week 1.
- Timeline: 6-12 weeks, mid-point checkpoint.
- Commercials: €0 to the institution if grant-funded (DRTIO next cycle / EUDIS); indicative production band Tier-1 €120-240k → Tier-3 €2.4-4.8m only if asked.
- Roles: Nexus Synergy (COP) + MAHI (edge) + sponsor (data + success criteria).
- After: an LOI if criteria are met.
- Funding vehicle: Research-Ireland Defence Innovation Challenge (next MDA-fusion-themed cycle) / EUDIS-EDF 2026 maritime track.
Stage 5 — LOI (the milestone — triggers Patrick's check). Trigger: pilot success criteria met. Owner: Buyer + Patrick. Activity: sign a non-binding LOI/MOU (founder signs in own name pre-incorporation, s.45 ratification; solicitor review required). Exit: signed LOI naming a sponsor + next step. Deliverable: LOI per checklist — parties; statement of intent; scope; success criteria referenced; named sponsor; indicative timeline; indicative commercials ("to be scoped under [vehicle]"); conditions; binding/non-binding stated; confidentiality; Irish governing law; Declined-List affirmation; signatures; solicitor review confirmed.
Stage 6 — Signed contract (paid pilot → framework). Trigger: LOI + a partner vehicle able to sign (MAHI prime / incorporated NewCo on Stamp 4). Owner: MAHI/NewCo + buyer procurement. Activity: convert to a paid pilot; line up the NMSS framework + EDF consortium for scale. Exit: a signed paid engagement. Deliverable: executed contract via the partner vehicle.
§16 — PM / timeline
Stage \ Month M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12
0 Verify(MAHI) ####
1 First contact ####
2 Free discovery ####
3 Scoped demo #####
4 Pilot (grant) ##########
5 LOI #### <-- Patrick-check trigger
6 Paid pilot/contract ############### (needs Stamp-1G/partner-vehicle gate)
CRITICAL PATH: [MAHI prime vehicle confirmed] ==========================> [PAID pilot signable]
Milestones: M1 MAHI bundle terms + named target · M2 discovery booked · M3 discovery write-up + demo agreed · M4-M5 demo delivered · M6-M9 pilot runs (grant-funded, €0 to buyer) · ~M9 LOI signed (Patrick check) · M10-M12 paid pilot via partner vehicle.
Critical path: the Stamp-1G / partner-vehicle gate is critical for a PAID pilot. Free discovery, demo and a grant-funded pilot can proceed without it; no founder-signed paid contract is possible until either MAHI is the contracting prime (V→0.75) or the founder reaches Stamp 4 and incorporates (V→1.0). t_resolve ≈ 9 months.
RACI:
| Activity | Founder (Don) | Patrick (sponsor) | MAHI (vehicle) | Buyer champion | Buyer procurement |
|---|---|---|---|---|---|
| Verify bundle terms | A/R | I | C | — | — |
| Warm intro to NS/RTI | C | A/R | I | I | — |
| Free discovery | A/R | C | C | R | I |
| Scoped demo | A/R | I | R | C | I |
| Pilot (grant-funded) | A/R | I | R | C | C |
| LOI | R | C (check) | C | A/R | C |
| Paid contract | C | I | A/R | C | A/R |
§17 — Funding / procurement vehicle
| Vehicle | Fit | Reality check [verified] |
|---|---|---|
| Research-Ireland Defence Innovation Challenge (next MDA-fusion cycle) | The named seed wedge. | The MSA slot is already won by Sea-Scan (Feb 2026, ~€1.8m) — a sensor, a feed into our COP, not a rival. Target the next cycle / a fusion-decision-layer-themed challenge; do not bank the current grant. |
| EUDIS / EDF 2026 | Strong, on-theme, non-dilutive. | EDF 2026 has an over-the-horizon maritime sensing call (~€56m) + a layered seabed-infrastructure-protection category + EUDIS SME calls; Ireland is EDF-eligible. Constraint: needs a ≥3-Member-State consortium (Ireland + Belgium/MAHI + one more). EDF 2026 closes 29 Sep 2026. Longer horizon, consortium-build effort. |
| National defence capital (NMSS 2026-2030) | The real prize, slower. | Ireland buys major defence via G2G / etenders.gov.ie, prefers European(-NATO) suppliers, neutrality-sensitive; a "national maritime security centre" software buy is a tendered framework — years at full scale, but a paid pilot can precede it. |
| IRCG (Dept of Transport) | Faster civil wedge. | Separate budget; a SAR-coordination pilot is lower-classification and may move quicker than the Defence line. |
Sequencing: civil/innovation pilot (IRCG SAR-COP or a DRTIO MDA-fusion pilot) → Naval Service paid pilot → NMSS framework + EDF consortium scale-up. Note: Ireland is excluded from NATO DIANA (non-NATO) — use EU vehicles only. [verified]
§18 — TWO-STAGE FORMULA SCORECARD
Every factor cell is
[PRIOR]with a dated one-line rationale (2026-06-05). Stage-1 weights {.22,.22,.18,.14,.10,.08,.06}; Stage-2 weights {.24,.22,.18,.16,.12,.08}.
Stage-1 — P(LOI) (vector {mandate_pull, access_warmth, demonstrability, decline_safety, white_space, cycle_speed, pillar_fit}):
| Factor | Score | Dated rationale (2026-06-05) [PRIOR] |
|---|---|---|
| mandate_pull | 4 | NMSS 2026-2030 (Feb 2026) explicitly proposes the national maritime security centre + undersea-infra protection; €60m sonar + cable-research programme funded. Not a 5: the centre is a concept, no dedicated COP budget line named yet. |
| access_warmth | 2 | Patrick's Defence-Forces network + Don's MAHI relationship are real, but no NS/RTI named contact is confirmed; the centre pilot-budget owner is unknown. |
| demonstrability | 3 | Storm/RMP pattern + AIS-multiplexer AS-IMPLEMENTED, but the MAHI data-contract spike is unbuilt and the spoofing Skill is AS-DESIGNED — a demo needs a real integration spike first. |
| decline_safety | 5 | Pure civil-security MDA / cable protection / SAR; clear of every Declined-List item; crew-ID forbidden by a DB CHECK; neutrality-compatible. |
| white_space | 2 | No sovereign-COP incumbent today, but Thales/Airbus prime gravity is a live lock-out risk (they could push their own fusion software downward). |
| cycle_speed | 1 | Full national-centre framework is G2G/OJEU, neutrality-sensitive — years; only a grant/civil pilot moves in months. |
| pillar_fit | 5 | This is the action-loop COP core — fusion + human-gated tasking over an editable maritime ontology, exactly our centre of gravity. |
S1 = 3.14 → P_LOI = 0.55 / (1 + exp(−1.15·(3.14 − 2.6))) = 35.8% [PRIOR].
Score100 = S1 · 20 = 62.8.
Reconciliation: board_anchor Score100 = 66.4 → 62.8 is within ~4 pts (acceptable). The raw model P_LOI (35.8%) runs above the board's to-LOI 18-30% band; that band is the gov-slow-tempered figure (G2G procurement, no named contact yet), so the reportable to-LOI = 18-30% stands as the calibrated value and the raw 35.8% is shown for transparency, not as the forecast.
Stage-2 — P(Contract | LOI) (vector {contractability, funding_to_pay, procurement_clarity, incumbency_displacement, time_to_value, reference_leverage}):
| Factor | Score | Dated rationale (2026-06-05) [PRIOR] |
|---|---|---|
| contractability | 3 | MAHI as contracting prime is plausible (founder-to-founder) but unconfirmed; the founder is solo on Stamp 1G. |
| funding_to_pay | 3 | EDF/EUDIS maritime calls + NMSS capital + IRCG line are real, but no named self-funded buyer line and EDF needs a ≥3-state consortium. |
| procurement_clarity | 2 | Sub-OJEU grant/innovation pilot route exists; the framework path is G2G/neutrality-sensitive and unclear. |
| incumbency_displacement | 3 | We sit above sensors (integrate, don't displace), but Thales/Airbus could bundle a competing fusion layer into a hardware framework. |
| time_to_value | 2 | A grant pilot can show value in a quarter, but full-value (the national picture) outruns most near-term windows. |
| reference_leverage | 2 | No delivered decline-safe reference yet; this is an early logo — a maritime win would later de-risk others. |
S2 = 2.62 → P_raw = 0.70 / (1 + exp(−1.1·(2.62 − 2.8))) = 31.5% [PRIOR].
Legal gate G = V·T, with T = min(1, max(0, (30 − 9)/30)) = 0.70 (t_resolve ≈ 9 mo):
| V (vehicle) | G = V·T | P(Contract|LOI) = G·P_raw | P(Contract) = P_LOI·G·P_raw |
|---|---|---|---|
| 0.10 (solo, Stamp 1G — today) | 0.07 | 2.2% | 0.8% |
| 0.45 (Dogpatch-brokered entity) | 0.315 | 9.9% | 3.6% |
| 0.75 (MAHI confirmed prime) | 0.525 | 16.6% | 5.9% |
| 1.00 (Stamp 4 + incorporated) | 0.70 | 22.1% | 7.9% |
Reconciliation to board to-buy 21-29%: the board's to-purchase band is the V≈0.75 (MAHI prime) regime — P(Contract|LOI) ≈ 16.6% at V=0.75 is the decisive lever; the board figure folds in μ_strategic/expansion upside (very-high value, national-centre + EDF scale). MAHI's value is that it lifts the gate (V 0.10→0.75), not Stage 1 — exactly worked-example (D). Today, solo, the honest joint number is ~0.8%.
§19 — Commercial
| Tier | Scope | ACV band | Terms / requirements |
|---|---|---|---|
| Tier-1 | One cable-corridor watch OR IRCG SAR-COP pilot | €120-240k/yr | Often €0 to the institution if grant-funded (DRTIO/EUDIS); EU-hosted; DPA + mutual NDA; partner vehicle (MAHI) to invoice. |
| Tier-2 | Multi-corridor EEZ RMP + SAR coordination, classification-partitioned | €500k-1.2m/yr | Paid pilot → framework; sovereign hosting (SecNumCloud-tier); FRIA + provenance ledger as deliverables. |
| Tier-3 | The "national maritime security centre" in software (whole-of-government COP, multi-feed) | €2.4-4.8m/yr | Tendered framework via etenders/G2G; multi-year; neutrality-compliant; EU-sovereign model routing + audit accreditation. |
Cost advantage: open-source substrate + sovereign HPC (EuroHPC LUMI / CASPIr-ICHEC) + a shared conformity file + no forward-deployed-engineer dependency (declarative composition). Indicative production pricing quoted only if asked; pilots lead with the grant route so the buyer isn't blocked on a procurement they can't yet run.
§20 — Legal blockers
| # | Blocker | Effect | Mitigation / status |
|---|---|---|---|
| X1 | Stamp 1G (founder) | No founder-signed paid contract until Stamp 4 — the critical-path gate for §16. | Contract via MAHI vehicle (V→0.75) now; or EEA-resident director (Patrick/Manuel) + incorporation; pursue Stamp 4. ⚖️ CONFIRM (immigration solicitor). |
| X2 | EU AI Act Annex-III high-risk + Art-27 FRIA | Maritime-surveillance COP is law-enforcement-adjacent high-risk → FRIA required; Art-5 prohibitions banned. | Conformity-by-construction; FRIA workflow; Declined-List filter; Art 5 applicable 2 Feb 2025. Per-use-case ⚖️ CONFIRM. |
| X3 | GDPR / data residency | If any crew personal data is touched, controller/processor duties + EU residency. | Designed-out: vessel telemetry only; crew-ID forbidden by DB CHECK; EU-hosted; DPA per buyer. [verified] |
| X4 | Dual-use export Reg (EU) 2021/821 | Vessel/comms monitoring can classify as controlled cyber-surveillance. | Export classification per integrated config; end-use/end-user due diligence. ⚖️ CONFIRM (export counsel). |
| X5 | Defence procurement (Dir 2009/81/EC + Art-346 TFEU) | A security-sensitive maritime contract can be exempted from open tender (helps or excludes). | Anticipate a negotiated/exempt procedure; get into the pilot conversation before a framework is specified. [verified] |
| X6 | NATO DIANA exclusion (Ireland non-NATO) | Irish entity ineligible for DIANA funding. | Use EU vehicles (EDF/EUDIS/PESCO) only. [verified] |
| X7 | Pre-incorporation contracting | Founder personally liable until NewCo formed. | Sign NDAs/MOUs in own name (s.45 ratifiable); solicitor-reviewed templates. ⚖️ CONFIRM. |
| Buyer-specific | Thales/Airbus prime lock-out + MAHI exclusivity unknown | A prime could bundle a fusion layer; MAHI may carry an Ultranav/edge exclusivity. | Position as the neutral fusion altitude; verify MAHI bundle terms via Don↔Pieter-Jan before any joint pitch. [unverified] |
§21 — Warm-intro contact + the specific ask
- Primary (founder-to-founder): Pieter-Jan Note (CEO, MAHI), via Don's direct relationship. The ask: a technical session on the MAHI Sense open-API / track schema to scope the data contract, and an honest conversation about MAHI acting as the contracting prime for a first Irish discovery/pilot (this lifts the legal gate V 0.10→0.75 — the single highest-leverage move in this package).
[verified person; email not public] - Bridge (institutional): Patrick Walsh (CEO, Dogpatch Labs),
patrick@dogpatchlabs.com. The ask: one named introduction to the Defence Forces RTI Unit (via the DEC) — ideally whoever is mapping the NMSS national maritime security centre — for a free, civil-security MDA discovery session (explicitly not warfighting).[verified] - Insider to map (do not cold-contact): Cdr Cathal Power (DF Liaison, Sea-Scan) — a known MDA-innovation insider; map, don't assume a relationship.
[verified name]
§22 — Open questions + consolidated Sources
Open questions (do not assert until resolved):
- Named buyer contact — exactly who in NS/RTI owns the centre pilot budget, and can Patrick reach them?
[TBD] - Pilot-budget existence — is there a direct innovation/pilot budget short of the full NMSS framework, or must everything route through a grant/tender?
[TBD] - MAHI bundle terms — edge scope, open-API reality, any Ultranav exclusivity.
[unverified] - Next Defence Innovation Challenge cycle — when, and is a fusion/decision-layer topic plausible vs the awarded sensor topics?
[TBD] - EDF consortium — can Ireland + Belgium/MAHI + a third Member State be assembled before the 29 Sep 2026 close?
[TBD] - Sovereignty bar — what classification/data-residency/audit requirements will the Defence Forces impose, and does our router + FRIA/provenance stack meet them as-built?
[TBD] - Civil-vs-defence split — is IRCG (Transport, SAR) a faster, cleaner first pilot than the Naval Service (Defence)? Likely yes — confirm.
[TBD]
Sources (web-verified 2026-06-04/05):
- NMSS 2026-2030 + national maritime security centre: gov.ie; RTÉ 25 Feb 2026; Naval Today 26 Feb 2026; RTÉ analysis 28 Feb 2026
- Naval Service capability/manpower: Irish Times 11 Jan 2025; Irish Examiner (strength)
- Sub-sea cables / EEZ threat: CSIS Ireland subsea case study; RTÉ Brainstorm 30 Jul 2025; Irish Times 14 Apr 2025 (intel network); Irish Times 5 Dec 2025 (cable research programme)
- €60m Thales sonar + "one company gains": Irish Times 15 Jun 2025; Irish Times 24 Jan 2026
- Sea-Scan (MSA winner, Cdr Cathal Power, DAS+SoP): TCD Feb 2026; Engineers Ireland
- IRCG Bristow SAR aviation: RTÉ 21 Aug 2025
- AI to cover DF shortfalls: Irish Times 16 May 2026
- EDF/EUDIS 2026 maritime + Ireland eligibility: EDF Work Programme 2026 (EC); Defence Forces European Funding page
- MAHI edge: mahi.be
- Internal grounding: source card
03-gtm/target-packages/maritime-naval-service.md; formula03-gtm/strategy-paper/formula-worked-examples.md(worked example D);discovery-playbook.md;outreach-drafts.md;legal-blockers-register.md;contact-register.md.
