Nexus Synergy Engagement Package — E08 · European Commission (EDF 2026 / EUDIS) consortium (min 3 Member States) — funded via a Belgium(MAHI) + Ireland(Nexus Synergy) + 3rd-MS team

STANDING BANNER (read first). Nexus Synergy Ltd is pre-incorporation, pre-revenue, pre-pilot. The European Commission / EDF-2026 call, EUDIS, and any consortium partners named here are a funding TARGET, not a customer or an award — nothing in this file is a grant, a deployment, an engagement, or traction; the ~178-surface gallery is a gallery, not a deployed system, and no consortium has been formed or has agreed to anything. Anti-cookie-cutter: the §8 mockups are programmatically generated from this opportunity's real entities (the EDF-2026 €1bn call, the CSIP €30M seabed-protection topic, the 29 Sep 2026 deadline, the WP1–WP5 consortium decomposition, the €120k EUDIS Business Accelerator voucher) — they are this opportunity's screens, not a reused template. Every probability below is [PRIOR] — a subjective pre-pilot estimate for prioritisation, not a forecast. Decline-safe / civil-security framing only: seabed-infrastructure protection (cables/pipelines/wind farms), maritime domain awareness, SAR and CISE — a protective command-and-control / decision-layer, never a weapons effector, never offensive targeting, never biometric ID. Ireland-neutrality-compatible by construction.

---

§0 — Header + one-line thesis

A sovereign, EU-only COP decision-layer contributed as one work-package (WP4) to a min-three-Member-State EDF-2026 consortium bidding the Layered Critical Seabed Infrastructure Protection (CSIP) research topic — the module that takes the consortium's fused underwater-sensor / USV-edge / AIS picture and runs the closed action-loop the topic literally calls "integrated command and control / maritime command centres": enrich → triage → human-gate → task → assess → close, fully audited (PROV-O), with a blocking 2-of-2 human sign-off so a protective response (alert a cable operator, re-task a USV to investigate, raise a CISE notification) is never automated. The play is non-dilutive: a 100%-funded EDF Research Action carries the build that every Irish-buyer package in this compendium then reuses.

§1 — Entity snapshot (cited)

• Funding authority (the "buyer"): the European Commission, DG Defence Industry and Space (DEFIS), which adopted the EDF 2026 Annual Work Programme on 17 December 2025 earmarking ~€1 billion for collaborative defence R&D across 10 calls / 31 call topics, with proposals closing 29 September 2026, 17:00 Brussels time. [verified — defence-industry-space.ec.europa.eu EDF-WP-2026; Commission press 17 Dec 2025]
• The instrument inside it: the EU Defence Innovation Scheme (EUDIS) — €231 million within the 2026 WP (≈€60M disruptive technologies; ≈€60M non-thematic SME calls), with SME research actions funded up to €4M (100% funded) and development up to €6M, plus a EUDIS Business Accelerator €120k seed voucher. [verified — eudis.europa.eu; zabala.eu EDF-2026]
• The topic we slot into: Layered Critical Seabed Infrastructure Protection (CSIP), a Research Action of ~€30M, open to consortia of ≥3 organisations from ≥3 different EU Member States or Norway, for a system-of-systems (unmanned assets, underwater sensors, detection/communication, integrated command and control, integrated with maritime command centres) to protect submarine cables, pipelines, wind farms, ports. [verified — ketmarket.eu / scientifyresearch.org CSIP topic; EDF-2026 Call Topic Descriptions]
• The contractable party: there is no single signing "customer." EDF funds a consortium (≥3 legal entities in ≥3 Member States). The realistic vehicle: MAHI BV (Belgium) as coordinator/prime on the edge, Nexus Synergy (Ireland) on the decision-layer WP, and a third Member-State partner (an underwater-sensor / command-centre house) — assembled around the EDF Partner Pool marketplace. [verified — consortium rule; partner-pool b2match; partner identity TBD]
• Context topic (NOT our slice): the EDF-2026 Enhanced Medium-Size Semi-Autonomous Surface Vessel (EMSAS) development topic (~€90M, littoral, ISR/ASuW/ASW/SBW modules) shows where the platform money sits; we are deliberately not a vessel programme — we are the protective C2/decision overlay. [verified — ketmarket.eu EMSAS]

§2 — The pain (web-verified, dated, cited)

The forcing function is two-sided and dated. (1) The threat: Europe's critical underwater infrastructure — the submarine cables and pipelines that carry its data and energy — has been repeatedly damaged since the 2024–25 Baltic cable-cut incidents, and the Commission's own CSIP topic text names the threat set explicitly (submarines, swimmer-delivery vehicles, combat divers, unmanned systems) and the gap: timely, robust detection, monitoring and management fused into integrated command and control with maritime command centres. (2) The money on a clock: the €1bn EDF-2026 call closes 29 September 2026 and the €30M CSIP research action is the named line for exactly this — a non-dilutive route a pre-revenue SME can win, but only inside a ≥3-Member-State consortium. [verified — EDF-2026 WP; CSIP topic]

The pain for Nexus Synergy specifically is the mirror image: it is pre-incorporation and pre-revenue, so every Irish public-sector pilot it wants (NDFEM, Naval Service, Coast Guard, SFPA, ComReg-subsea) is blocked on a legal vehicle (Stamp-1G) and on customer budget that cannot move quickly. A 100%-funded EDF Research Action inverts both blockers: the consortium (with MAHI as prime) is the vehicle, the Commission is the payer, and the deliverable — a sovereign COP decision-layer — is the same artefact every domestic package then reuses. The risk is the opposite of "no funding": it is prime gatekeeping — large EDF C2 work-packages are routinely captured by Thales / Naval Group / Indra / Leonardo, so a pre-revenue SME must enter as a specialised, well-scoped WP with a confirmed prime, not as a would-be system integrator. [verified — EDF participation norms; legal-blockers-register X6]

§3 — Use-case & value (DECLINE-SAFE framing)

What this is: a protective decision-layer for civil-security seabed-infrastructure protection inside an EU-funded research consortium. It ingests the consortium's underwater picture (distributed-acoustic-sensing on fibre, passive/active acoustic, USV/sensor-edge tracks via the MAHI Sense open API, AIS) and closes a human-gated protective action — notify a cable operator, re-task an uncrewed surface vessel to investigate an anomaly near a cable corridor, raise a CISE maritime-information-sharing notification, cue a SAR asset — with a tamper-evident audit trail.

What this is explicitly NOT (stated in the consortium Description-of-Action and in every showing): not a weapons effector or fire-control system; not offensive targeting; not real-time biometric identification; not emotion recognition, social scoring, untargeted scraping, psychometric profiling, or predictive policing. The product observes infrastructure and vessels and recommends a protective action to a human — it never authorises a kinetic effect, and the schema and the gate make that structural (see §5/§7). Because Ireland is militarily neutral (Lisbon Protocol guarantee), Nexus Synergy's WP is scoped to protective C2 only — a decision/audit overlay, not a platform or an effector — which is both the neutrality-compatible posture and the decline-safe line. [verified — Irish neutrality protocol; AI Act Art-5 applicable 2 Feb 2025]

Value to the consortium (and the Commission): a sovereign, conformity-by-construction decision-layer (EU-only model routing via the AI Provider Router; FRIA + PROV-O ledger as first-class components) that lifts the consortium's TRL on the command-and-control deliverable the topic scores, and gives the Commission a visibly auditable, human-in-the-loop answer to the dual-use governance question that always shadows EDF C2 proposals. Value to Nexus Synergy: non-dilutive funding for the exact build the rest of this compendium depends on, plus an EU-scale reference.

§4 — Ontology (this opportunity's domain entities + relationships)

ConsortiumPartner (legal entity; member_state; role=coordinator/beneficiary/affiliated)
    └─ DELIVERS → WorkPackage (WP1 edge / WP2 sensors / WP3 cmd-centre / WP4 COP / WP5 conformity)
WorkPackage
    └─ FUNDED_BY → CallTopic (CSIP RA; budget; deadline; min_MS=3)        ── EU grant line
CriticalUnderwaterInfra (cable / pipeline / wind-farm / port; corridor_geom; operator)
    └─ PROTECTED_BY → ProtectiveResponse                                  ── DEFENSIVE only
SensorObservation (DAS-fibre / acoustic / AIS / USV-edge track)          ── AGGREGATE/track
    └─ NEAR → CriticalUnderwaterInfra   (proximity-to-corridor is the trigger feature)
AnomalyCue (loiter / AIS-gap / acoustic-contact near a cable corridor)
    └─ DERIVED_FROM → SensorObservation (a derived cue, NEVER an auto-action)
ProtectiveResponse (notify_operator / re-task USV to investigate / CISE notify / cue SAR)
    └─ AUTHORISED_BY → DualControlSignoff (2-of-2 QES)
        └─ REQUIRES → (no kinetic effect; protective/observe-investigate-notify ONLY)
DecisionRoomVerdict (5 personas incl. civil-society + proportionality)
    └─ GATES → ProtectiveResponse   (LOAD-BEARING: no response without it)
ProvenanceRecord (PROV-O; wasAttributedTo agent; Merkle/Rekor anchor)
    └─ TRACES → every node above (audit-grade reconstruction for the Commission/MS)

The AnomalyCue is a derived entity surfaced to a human, never a fact and never an auto-tasker; ProtectiveResponse is constrained at the ontology to the protective verb set (notify / investigate / cue) — there is no offensive verb in the schema.

§5 — Data model (synergy.* tables; RLS + a load-bearing CHECK)

SQL
56 lines
Copy
-- Consortium binding: which partner owns which work-package, scoped per consortium tenant.
CREATE TABLE synergy.consortium_workpackage (
    id              uuid PRIMARY KEY DEFAULT gen_random_uuid(),
    org_id          uuid NOT NULL,                              -- RLS: app.current_org_id (= consortium tenant)
    partner_entity  text NOT NULL,                              -- e.g. MAHI BV / Nexus Synergy Ltd
    member_state    char(2) NOT NULL,                           -- ISO MS code (BE/IE/...)
    role            text NOT NULL CHECK (role IN ('coordinator','beneficiary','affiliated')),
    work_package    text NOT NULL CHECK (work_package IN ('wp1_edge','wp2_sensors',
                       'wp3_cmd_centre','wp4_cop_decision','wp5_conformity')),
    call_topic      text NOT NULL DEFAULT 'CSIP-RA',
    grant_share_eur numeric,                                    -- indicative EU share to this WP
    prov_source     text NOT NULL
);
ALTER TABLE synergy.consortium_workpackage ENABLE ROW LEVEL SECURITY;
CREATE POLICY org_isolation ON synergy.consortium_workpackage
    USING (org_id = current_setting('app.current_org_id')::uuid);

-- Sensor observations near a protected corridor. AGGREGATE/track-level; no person data.
CREATE TABLE synergy.seabed_observation (
    id              uuid PRIMARY KEY DEFAULT gen_random_uuid(),
    org_id          uuid NOT NULL,
    source          text NOT NULL CHECK (source IN ('das_fibre','acoustic','ais','usv_edge')),
    track_geom      geometry(Point, 4326) NOT NULL,
    corridor_id     uuid,                                       -- the CriticalUnderwaterInfra it is NEAR
    proximity_m     numeric CHECK (proximity_m >= 0),           -- distance to corridor (the trigger feature)
    person_identifier text                                      -- MUST stay NULL (decline-safe)
        CHECK (person_identifier IS NULL),                      -- LOAD-BEARING: no individual data, ever
    observed_at     timestamptz NOT NULL,
    prov_source     text NOT NULL                               -- PROV-O: sensor + snapshot id
);
ALTER TABLE synergy.seabed_observation ENABLE ROW LEVEL SECURITY;
CREATE POLICY org_isolation ON synergy.seabed_observation
    USING (org_id = current_setting('app.current_org_id')::uuid);

-- A PROTECTIVE response is a first-class audited entity. It is (a) restricted to a protective
-- verb set (NO kinetic/offensive effect) and (b) un-executable without a 2-of-2 human sign-off.
CREATE TABLE synergy.protective_response (
    id              uuid PRIMARY KEY DEFAULT gen_random_uuid(),
    org_id          uuid NOT NULL,
    corridor_id     uuid NOT NULL,
    kind            text NOT NULL CHECK (kind IN
                       ('notify_operator','retask_usv_investigate','cise_notify','cue_sar')),
    -- LOAD-BEARING (decline-safe): the verb set has NO offensive option; a kinetic effect
    -- cannot be represented in this table at all.
    proportionality numeric(3,2),
    sign_op         text,                                       -- consortium operational signer QES; NULL until signed
    sign_oversight  text,                                       -- proportionality/oversight QES;      NULL until signed
    executed_at     timestamptz,                                -- NULL until tasked
    merkle_leaf     bytea NOT NULL,
    prov_derived_from text NOT NULL,                            -- PROV-O: which observations + anomaly rule
    -- LOAD-BEARING INVARIANT: the human-gated decision. No execution without 2-of-2 sign-off.
    CHECK (executed_at IS NULL OR (sign_op IS NOT NULL AND sign_oversight IS NOT NULL))
);
ALTER TABLE synergy.protective_response ENABLE ROW LEVEL SECURITY;
CREATE POLICY org_isolation ON synergy.protective_response
    USING (org_id = current_setting('app.current_org_id')::uuid);

Three load-bearing invariants make the decline-safe promise structural: (1) seabed_observation.person_identifier IS NULL — the schema physically cannot store an individual; (2) protective_response.kind is constrained to a protective-only verb set — there is no offensive/kinetic option to insert; (3) protective_response cannot be executed_at without both QES signatures — the human gate lives in the database, not just the UI.

§6 — Action-loop pseudocode (detect → enrich → triage → gate → task → execute → BDA → close)

Python
44 lines
Copy
def protect_seabed_corridor(consortium_state, observations, corridor):
    # 1. DETECT — a sensor contact appears NEAR a protected cable/pipeline corridor.
    contacts = [o for o in observations if o.corridor_id == corridor.id and o.proximity_m < CORRIDOR_M]
    if not contacts:
        return Disposition.MONITOR_ONLY

    # 2. ENRICH — fuse the consortium picture: DAS-fibre + acoustic + AIS + USV-edge tracks.
    fused = fuse_multi_sensor(contacts, mahi_sense_api=consortium_state.edge_bus)   # WP1 edge feeds in
    analogue = cma_collective_memory.recall_analogue(corridor)                      # -> Baltic 2024-25 pattern

    # 3. TRIAGE — the protective cue: vessel loiter / AIS-gap / acoustic contact near the corridor.
    cues = [c for c in fused if c.is_loiter or c.ais_gap or c.acoustic_unattributed]
    if not cues:
        return Disposition.NO_ANOMALY
    top = rank_by(cues, key="proximity_and_persistence")[0]                          # the closest, most persistent

    # 4. GATE — human Decision Room; PROTECTIVE response only (no kinetic verb exists).
    gate = decision_room(
        proposal  = ProtectiveResponse(corridor, kind=propose_protective_action(top)),  # notify/investigate/cise/sar
        personas  = [OP_TASKING, PROPORTIONALITY, CIVIL_SOCIETY("AlgorithmWatch"),
                     DEVILS_ADVOCATE, DOMAIN_EXPERT("Coast-Guard/Cable-Operator")],
        threshold = P_MIN_TIER2)
    assert gate.proposal.kind in PROTECTIVE_VERBS                                    # belt-and-braces with the DB CHECK
    if gate.civil_society.flag == "ATTRIBUTION_UNCERTAIN":
        proposal.downgrade_to("retask_usv_investigate")                             # investigate, do not escalate
    if not gate.passed:
        audit.append(top.id, f"refused: {gate.rationale}", merkle=True)
        return Disposition.REFUSED(gate.rationale)

    # 5. TASK + 6. EXECUTE — only after 2-of-2 QES; the DB CHECK also forbids it otherwise.
    sign_op   = await qes_sign("consortium.operational", proposal)
    sign_ovr  = await qes_sign("consortium.oversight", proposal)
    if not (sign_op and sign_ovr):
        audit.append(top.id, "protective action unsigned -- withheld", merkle=True)
        return Disposition.WITHHELD
    receipts = execute([dispatch(proposal)])           # notify operator / re-task USV / CISE notify / cue SAR
    write_back("synergy.protective_response", proposal, executed_at=now(),
               sign_op=sign_op, sign_oversight=sign_ovr, merkle=True)

    # 7. BDA — did the USV reach the corridor / did the operator acknowledge / was the contact resolved?
    bda = assess_protective_outcome(receipts, corridor)

    # 8. CLOSE — UNO loop close; register a ProspectiveItem if a sensor-gap recurred (feeds WP5 conformity).
    return close_loop(top, bda, prospective=detect_recurring_sensor_gap(receipts))

§7 — nexus-workflows YAML DAG (same loop, declarative, blocking human-gate node)

YAML
38 lines
Copy
# nexus-workflows job: seabed_protection_cop_loop  (NO cron — triggered as a workflow)
name: seabed_protection_cop_loop
trigger:
  on: sensor_contact_near_corridor       # DAS/acoustic/AIS/USV-edge proximity event
  org: "{{ consortium.org_id }}"
nodes:
  - id: detect
    run: synergy.detect.contact_near_corridor
    with: { corridor_threshold_m: 2000 }
  - id: enrich
    needs: [detect]
    run: synergy.fuse.multi_sensor
    with: { sources: [das_fibre, acoustic, ais, usv_edge], edge_api: mahi_sense }
  - id: triage
    needs: [enrich]
    run: synergy.triage.protective_anomaly_cue        # loiter / AIS-gap / acoustic-unattributed
  - id: human_gate                        # BLOCKING — workflow halts here until a human signs
    needs: [triage]
    type: human_approval
    blocking: true
    decision_room:
      personas: [op_tasking, proportionality, civil_society_algowatch,
                 devils_advocate, domain_coastguard_cable_operator]
      threshold: P_MIN_TIER2
      verb_constraint: protective_only                # notify / investigate / cise / sar — NO kinetic
      dual_control: { signers: 2, method: QES }        # mirrors the DB CHECK on protective_response
    on_reject: { goto: close, status: refused }
  - id: task_execute
    needs: [human_gate]
    run: synergy.execute.protective_response
    with: { targets: [cable_operator, usv_tasking, cise_node, sar_cue] }
  - id: bda
    needs: [task_execute]
    run: synergy.assess.protective_outcome
  - id: close
    needs: [bda]
    run: synergy.uno.close_loop
    emits: prospective_item                            # recurring sensor-gap -> WP5 conformity backlog

The human_gate node is blocking and verb_constraint: protective_only — the DAG cannot reach task_execute until two named signers approve a protective action. This is the same pair of invariants the database enforces (the 2-of-2 CHECK and the protective-only kind constraint), expressed declaratively. It is the literal, auditable demonstration that a sovereign EU decision-layer can sit inside a defence-fund consortium without becoming an autonomous weapons system.

§8 — UI/UX mockups (VERBATIM generated ASCII)

Surface 1 (PRIMARY) — command_dashboard. Caption: ShellLayout + TopBar (with PccPill "CSIP / WP4 decision-layer" in rightSlot, RendererChip = board) + LeftSidebar glyph-rail + a KPI-tile row (call/deadline · our WP4 slice · the min-3-MS consortium) over the WP1–WP5 work-package decomposition feed and the gate/budget/milestone lines, with the right-docked ChatTerminal Decision Room + HistoryRail (PROV-O) + BottomStatusBar + ClassificationBanner (EU-RESTRICTED // CIVIL-SECURITY // NON-OFFENSIVE).

Figure — Command dashboard (KPI + live feed). Production-fidelity React surface (buildable); the faithful ASCII follows.

+--------------------------------------------------------------------------------------------------+
| EDF-2026 CSIP Consortium -- Seabed-Infra-Protection COP decision-layer (command_dashboard)       |
+--------------------------------------------------------------------------------------------------+
| CALL / DEADLINE                | OUR SLICE (WP 4)               | CONSORTIUM (min 3 MS)          |
| EDF-2026 (EUR 1bn)             | COP decision-layer C2          | BE + IE + (3rd MS TBD)         |
| closes 29 Sep 2026 17:00       | in CSIP RA, EUR 30M topic      | EU-only / EDIP-compatible      |
+--------------------------------------------------------------------------------------------------+
| LIVE ALERT FEED                                                                                  |
|  WP1 MAHI(BE)  USV/sensor edge -- Alpha-8 + MAHI Sense open API track bus                        |
|  WP2 PARTNER   underwater sensors (DAS fibre, passive/active acoustic)                           |
|  WP3 PARTNER   maritime command centre integration / S-100 ENC base                              |
|  WP4 NEXUS(IE) the SOVEREIGN COP DECISION-LAYER -- enrich/triage/gate/BDA                        |
|  WP5 ALL       conformity file: FRIA + PROV-O ledger + AI-Provider-Router                        |
|  GATE  partner-vehicle: Nexus contributes via MAHI prime until Stamp-4 (V 0.10->0.75)            |
|  BUDGET ask EUR 0.8-1.4M to Nexus WP4 (100% RA-funded; non-dilutive grant)                       |
|  MILESTONE  consortium LoI bundle + Part-B draft locked by 15 Aug 2026                           |
+--------------------------------------------------------------------------------------------------+

Surface 2 — settings_form. Caption: the consortium-binding + WP4 COP decision-layer module configuration rendered as a SkillBindingInspector settings surface — three sections (EDF eligibility binding; the WP4 module's input bus / loop / immutable human-gate / EU-only AI Provider Router / decline-safe scope / hard exclusions / conformity; and the sovereignty + Irish-neutrality posture). This is the declarative artefact a reviewer (or the Commission) reads to confirm the decision-layer is protective, human-gated, EU-sovereign, and neutrality-compatible.

Figure — Integration / binding form. Production-fidelity React surface (buildable); the faithful ASCII follows.

+--------------------------------------------------------------------------------------+
| Consortium Binding + WP4 COP Decision-Layer (SkillBindingInspector)                  |
+--------------------------------------------------------------------------------------+
| Consortium binding (EDF eligibility)                                                 |
| ------------------------------------                                                 |
|   Call topic                  CSIP -- Layered Seabed Infra Protect                   |
|   Action type                 Research Action (RA), 100% funded                      |
|   Min eligibility             >=3 legal entities / >=3 Member States                 |
|   Coordinator                 MAHI BV (Belgium) -- prime/vehicle                     |
|   WP4 beneficiary             Nexus Synergy (IE) -- decision-layer                   |
|   3rd Member State            TBD (NL/ES/EE partner -- discovery)                    |
|   Funding to Nexus            EUR 0.8-1.4M [TBD on WP scope]                         |
|   Deadline                    29 Sep 2026 17:00 Brussels                             |
|                                                                                      |
| WP4 COP decision-layer module (binding)                                              |
| ---------------------------------------                                              |
|   Input bus                   MAHI Sense API + DAS/acoustic + AIS                    |
|   Loop stages                 enrich->triage->gate->task->BDA->close                 |
|   Human gate                  blocking 2-of-2 QES (ON, immutable)                    |
|   AI Provider Router          EU-only; refuse out-of-jurisdiction                    |
|   Decline-safe scope          seabed-infra + MDA + SAR + CISE                        |
|   Excluded (hard)             no weapon effect / no biometric ID                     |
|   Provenance                  PROV-O ledger + Merkle/Rekor anchor                    |
|   Conformity                  FRIA (AI Act Annex-III) shared file                    |
|                                                                                      |
| Sovereignty + neutrality posture                                                     |
| --------------------------------                                                     |
|   Data residency              EU-hosted; no US hyperscaler model call                |
|   IE neutrality               protective C2 only -- not an effector                  |
|   Export (2021/821)           dual-use class per config [CONFIRM]                    |
|   IP licence chain            Adverant->Nexus one-way perpetual                      |
|   Background IP               MAHI edge carved out (pre-existing)                    |
+--------------------------------------------------------------------------------------+

(Bespoke SVG of the primary screen: _build/figures/edf2026-eudis-consortium/uc-command_dashboard.svg.)

§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

Literal click-path: The founder opens the command_dashboard consortium board → reads the WP1–WP5 decomposition and confirms WP4 is the in-scope slice (PccPill shows "CSIP / WP4 decision-layer") → clicks the WP4 feed row to open the settings_form SkillBindingInspector → verifies the Human gate = blocking 2-of-2 QES (ON, immutable), the AI Provider Router = EU-only, the Excluded (hard) line (no weapon effect / no biometric ID), and the IE neutrality = protective C2 only posture → opens the right-docked ChatTerminal Decision Room to run a synthetic seabed-anomaly through the gate (proving the protective-only verb set) → clicks [ Bind to WP4 ] → the binding, the PROV-O chain, and the milestone (LoI bundle locked 15 Aug) land on the HistoryRail.

[command_dashboard board] -> [click WP4 feed row] -> [settings_form SkillBindingInspector]
        -> [verify gate ON / EU-only router / hard exclusions / neutrality posture]
        -> [ChatTerminal Decision Room: run synthetic anomaly through protective-only gate]
        -> [Bind to WP4] -> [HistoryRail: binding + PROV-O + milestone]

§10 — Decision-Room transcript (the gated decision)

Scenario: during a CSIP demo, the fused picture shows an unattributed acoustic contact and an AIS-gap vessel loitering ~1.2 km from a submarine-cable corridor off the consortium's demo coastline. The Decision Room must decide a protective response.

Op-tasking · L3 EuroLLM-EU: "Recommend: notify the cable operator and re-task the nearest USV to investigate the loiter. Both are protective/observe actions. Persistence 18 min, proximity 1.2 km, AIS dark since 06:51. This is exactly the 'integrated command and control' deliverable the CSIP topic scores."

Proportionality · L3 Mistral-Large-EU: "Lawful basis: protection of critical infrastructure; data-minimised — we act on track-level observations and a proximity feature, not on any individual. Necessity/proportionality/legitimacy P=0.88 [N:0.93 P:0.88 L:0.84]. Proportionate for investigate + notify. I would not support any escalation beyond observe — and the schema gives me none to support."

Civil-society reviewer · AlgorithmWatch: "Two conditions. (1) Attribution is uncertain — a fishing vessel loitering over a wreck looks identical to a hostile survey. Down-grade from any 'escalate' framing to investigate-only until the USV returns ground-truth. (2) The consortium Description-of-Action must publish the Declined-List clause — protective C2 only, no kinetic effect, no biometric ID — so the Commission and the public can hold the system to it. Add both or I dissent." → both conditions accepted; action set to retask_usv_investigate + notify_operator.

Devil's advocate · L2 Mistral-7B (local): "Is this just a fishing boat — are we crying wolf? Cross-check: DAS-fibre strain and passive acoustic both register the contact; this is not a single-sensor artefact. And the chosen action is the cheapest, most reversible one — send a USV to look. Holding does nothing while a cable corridor sits unobserved. Hold = NO."

Domain expert · Coast-Guard / Cable-Operator liaison: "The corridor is a live data cable; the operator wants a heads-up on any loiter inside 2 km regardless of attribution. USV is 9 minutes out. Concur — notify + investigate, with the civil-society conditions carried. And log that we did not escalate — that record is the product."

GATE VERDICT — ACCEPT (conditional, protective-only): P=0.88 ≥ P_min; action constrained to retask_usv_investigate + notify_operator (no kinetic verb exists in the schema); conditions carried (investigate-only until ground-truth; publish the Declined-List clause in the DoA); 2-of-2 dual-control QES required before any execution. Dissent log: AlgorithmWatch's attribution-uncertainty downgrade is recorded as a discoverable provenance edge — if the contact later proves benign, the chain proves the system chose restraint, which is the neutrality-and-decline-safety proof the EDF reviewer is looking for.

§11 — Write-back + BDA + PROV-O chain + deltas-only regulatory traceback

Write-back. On 2-of-2 sign-off, a WriteBackBubble in the ChatTerminal posts the protective_response (retask_usv_investigate + notify_operator) with executed_at=now(), Merkle-anchored. Tasking is written back to the USV via the MAHI Sense open API (WP1 owns the edge); the operator notification and any CISE message are written to the consortium's command-centre integration (WP3).

BDA (loop close). Outcome is assessed against the USV's return: contact resolved as a benign survey vessel at T+14m; operator acknowledged the notification at T+3m. A residual sensor-gap (the AIS-gap detection latency) registers a ProspectiveItem to the WP5 conformity backlog — closing the loop and feeding the consortium's continuous-improvement record.

PROV-O attribution chain (every node carries prov:wasAttributedTo agent + prov:wasGeneratedBy activity; Merkle-anchored to a Sigstore-Rekor-style transparency log):

DAS-Strain(Entity, fibre sensor)  +  AcousticContact(Entity, passive array)  +  AIS-Gap(Entity, AIS feed)
        └─ used ─► Multi-Sensor-Fusion(Activity)  ── wasInformedBy ── USV-EdgeTrack(Entity, MAHI Sense)
                          └─ wasGeneratedBy ─► AnomalyCue:loiter-near-corridor(Entity)
                                  └─ used ─► DecisionRoom-CSIP-DEMO(Activity, 5 personas + CG/operator)
                                          └─ wasAssociatedWith ─► DualControlSigners(Agent, Op+Oversight 2/2)
                                                  └─ wasGeneratedBy ─► ProtectiveResponse:investigate+notify(Entity)
                                                          └─ used ─► USV-Retask + Operator-Notify(Activity)
                                                                  └─ BDA: contact benign T+14m (Obs-L4) ─► UNO-close

Regulatory traceback (deltas only — what differs from the platform defaults). EU AI Act: a protective-C2 decision-support system in security is treated as high-risk ⇒ Art 27 FRIA completed for the COP pipeline (it is the shared WP5 conformity file); Art 14 human-oversight is satisfied by the blocking 2-of-2 gate; Art 5 prohibited practices (applicable 2 February 2025) are excluded by the protective-only verb set and the person_identifier IS NULL schema. EDF-specific delta: Art-12(6) EDF ethics review + the dual-use export check (Reg (EU) 2021/821) on the integrated configuration — flagged ⚖️ CONFIRM. Neutrality delta: Nexus Synergy's WP is scoped to protective C2 only, never an effector, which is the Irish-neutrality-compatible posture recorded in the DoA. GDPR delta: track-level/aggregate only — no personal-data processing in the WP4 module.

§12 — Buyer & stakeholders

§13 — Competition / incumbency + comparator

Incumbency in EDF C2 work-packages: the command-and-control slice of large EDF maritime/seabed consortia is the home turf of Thales, Naval Group, Indra, Leonardo and similar primes — they routinely lead the integration WP. This is the central risk: a pre-revenue SME cannot lead a €30M consortium's C2 integration and should not try. [verified — EDF participation norms; legal-blockers X6 / contact-register E]

The white space (and the strategy): what the primes do not bring is a sovereign, conformity-by-construction, human-gated decision-layer with a published Declined-List and EU-only model routing — a governance-grade C2 overlay, not a classical mission-system. Strategy: enter as a specialised, well-scoped WP under a confirmed prime (MAHI), delivering the auditable decision/PROV-O layer the topic's "integrated command and control" language scores and the Commission's ethics review rewards — integrate alongside the primes' systems, never displace them. Comparator framing (folded): versus a Palantir-class C2 product, the differentiators are structural sovereignty (AI Provider Router refuses out-of-jurisdiction calls), conformity-by-construction (FRIA + PROV-O + transparency as first-class), and the seated civil-society reviewer — the exact axes an EDF dual-use ethics panel probes. Honest loss: primes bring decades of accredited maritime-C2 deployment and existing security clearances; Nexus Synergy brings neither yet — which is precisely why the WP must be a scoped decision-layer, not the integration spine.

§14 — Readiness (honest, pre-pilot)

Reusable today (on /synergy-console): MapConsole triple renderer incl. the CesiumJS globe RMP; the COP entity model + /api/v1/cop/picture (117 live entities incl. DOLPHIN-21, isolated adverant-demo tenant); HistoryRail bitemporal playback + audit; the ChatTerminal right-dock Decision Room; the EM/SIGINT/CUAS/COP routes already merged on synergy-server; the AI Provider Router concept (EU-only model routing); the FRIA / PROV-O conformity components.

Real gaps (build, honestly): the CSIP seabed-fusion adapters (DAS-fibre strain, passive/active acoustic, AIS, USV-edge via the MAHI Sense open API — the data-contract spike is the first real engineering dependency, owned with MAHI's Bertold Van den Bergh / Quinten Lauwers); the proximity-to-corridor trigger and protective-only verb set; the synthetic seabed-anomaly demo dataset; the WP5 shared conformity file (FRIA for a defence-C2 context — get it right, do it with counsel). Non-technical critical gap: there is no consortium yet — assembling a ≥3-MS team with a confirmed prime and a 3rd partner by the 29 Sep deadline is the dominant risk, not code. Maturity statement for every showing: demo-ready (faithful gallery + globe RMP + a synthetic seabed scenario), not deploy-ready; no live tasking against any real asset, no formed consortium, no submitted proposal.

§15 — ENGAGEMENT PLAYBOOK

Warm-intro / first-contact note (adapt from outreach-drafts §2, to Pieter-Jan Note):

Subject: Edge + sovereign brain on the CSIP call — an all-EU Belgium+Ireland consortium for 29 Sep Pieter-Jan — the EDF-2026 Layered Critical Seabed Infrastructure Protection research action (€30M, closes 29 Sep 2026) is the cleanest non-dilutive vehicle for our edge+brain thesis: MAHI owns WP1 (the USV/sensor edge), Nexus Synergy owns WP4 (the sovereign COP decision-layer) — the "integrated command and control / maritime command centre" deliverable the topic literally scores. An all-EU Belgium+Ireland team is exactly what EDIP's buy-European mandate rewards, and a 100%-funded Research Action means neither of us fronts the build. Two asks: (1) would MAHI act as coordinator/prime? (2) a technical session on the MAHI Sense open API / track schema to scope the WP1↔WP4 data contract. Candidly, MAHI-as-prime also clears a legal/structuring gate on my side. — [Founder]

60-min discovery agenda (adapted for a consortium partner, from discovery-playbook §3): 0:00 frame ("scoping a joint WP, not selling"); 0:05 walk the CSIP topic text — which deliverable does each of us own; 0:20 the data-contract spike (MAHI Sense API ↔ WP4 COP bus) + the protective-only verb constraint; 0:35 the 3rd-MS partner (sensors/command-centre) — who, via the EDF Partner Pool; 0:45 the conformity file (who owns FRIA, ethics review, export classification); 0:55 lock the consortium-shape memo + the 15 Aug Part-B-lock date.

1-page WP4 proposal (filled for THIS opportunity): Title: WP4 — Sovereign COP Decision-Layer for Critical Seabed Infrastructure Protection. The one thing we'll prove: "A human-gated, EU-sovereign decision-layer can fuse the consortium's DAS/acoustic/AIS/USV-edge picture and close a protective action-loop (notify/investigate/CISE/SAR) with audit-grade PROV-O reconstruction and zero offensive capability." Scope IN: the WP4 module, the enrich→triage→gate→BDA loop, FRIA/PROV-O conformity (with WP5). Scope OUT / Declined: any kinetic/offensive effect, biometric ID, predictive analytics, autonomous tasking. Data & lawful basis: track-level/aggregate, EU-hosted, AI-Provider-Router EU-only. Success criteria: (1) protective loop closes with 2-of-2 gate in the demo; (2) PROV-O reconstruction passes an ethics-review tabletop; (3) attribution-uncertain cases downgrade to investigate-only, on the record. Timeline: proposal build to 29 Sep 2026; project ~24–36 mo post-award. Commercials: €0 to Nexus' counterparties — 100% EU-funded Research Action; indicative WP4 share €0.8–1.4M [TBD]. After: consortium commitment letter → (if awarded) Grant-Agreement accession.

LOI / consortium-commitment checklist (key items): parties (MAHI prime + Nexus Synergy, founder in own name pending NewCo, s.45 ratification); statement of intent (participate as WP4 beneficiary, conditional on award); WP scope + budget share carried from the WP4 proposal; Declined-List affirmation (protective C2 only; no kinetic/biometric); background-IP carve-out (MAHI edge); non-binding save confidentiality; governing law (Belgian for the consortium agreement, Irish for Nexus' internal arrangements — ⚖️ CONFIRM); solicitor review confirmed before signing.

§16 — PM / timeline

Stage              Jun  Jul  Aug  Sep  | (award decision lags)  Q1'27  Q2'27  ...  Y2-Y3
First contact(MAHI) ##
Discovery/scoping     ####
Consortium assembly     #########       (3rd-MS partner via EDF Partner Pool)
Scoped WP4 demo            ####
Part-B / DoA write           #######
LoI / commitment bundle        ####  (LOCK 15 Aug)
SUBMIT (29 Sep)                    ^submit
Vehicle/Stamp       ==============================>  (Stamp-1G -> Stamp-4 / MAHI-prime: CRITICAL for a PAID/funded WP)
Evaluation (Commission)              ............ (several months) ............
Grant Agreement                                                    ####
WP4 delivery                                                          ###############>

Milestones: Jun MAHI scoping note · Jul consortium shape agreed · Aug scoped WP4 demo + Part-B draft · 15 Aug LoI/commitment bundle locked · 29 Sep 2026 proposal submitted · ~Q1–Q2'27 evaluation result · post-award Grant-Agreement accession (the funded "contract"). Critical path: NOT the code — it is (a) the legal vehicle (a solo Stamp-1G founder cannot be a beneficiary that signs/invoices; MAHI-as-prime carries it, lifting the gate V from 0.10 to 0.75) and (b) assembling a confirmed ≥3-MS consortium before 29 Sep. Both must resolve for a funded WP; the demo and a non-binding commitment letter can proceed before they do.

RACI:

§17 — Funding / procurement vehicle

1. EDF-2026 CSIP Research Action (~€30M topic; 100%-funded; deadline 29 Sep 2026) — the primary vehicle; Nexus Synergy participates as a WP4 beneficiary inside a ≥3-MS consortium. Non-dilutive; no customer budget required. [verified]
2. EUDIS SME track (research up to €4M / development up to €6M; €231M EUDIS envelope) — alternate/complementary vehicle if a leaner SME-led consortium is faster to assemble than the full CSIP topic. [verified]
3. EUDIS Business Accelerator €120k seed voucher — a separate, lower-threshold non-dilutive step the founder can pursue in parallel to mature the WP4 module. [verified]
4. Procurement route: there is no OJEU here — the route is a competitive grant call (EDF eval) → Grant Agreement accession; the prime (MAHI) signs the GA and Nexus accedes as a beneficiary. Hard constraint: DIANA (NATO) is excluded because Ireland is non-NATO — EU vehicles only. [verified — diana.nato.int; legal-blockers X6]

§18 — TWO-STAGE FORMULA SCORECARD

Stage-1 (P_LOI) — S1 = Σ wᵢ·fᵢ; P_LOI = 0.55/(1+exp(-1.15·(S1-2.6))). Every factor [PRIOR], dated 2026-06-05.

S1 = 5(.22)+3(.22)+3(.18)+4(.14)+3(.10)+3(.08)+5(.06) = 3.70 → P_LOI ≈ 0.43 (band ~38–48%). ✓

Stage-2 (P_raw) — S2 = Σ vⱼ·gⱼ; P_raw = 0.70/(1+exp(-1.1·(S2-2.8))).

S2 = 2(.24)+5(.22)+4(.18)+3(.16)+2(.12)+2(.08) = 3.18 → P_raw ≈ 0.42.

Legal gate G = V·T (H_conv=30mo, t_resolve=10mo ⇒ T = (30-10)/30 = 0.667):

• V=0.10 (solo Stamp-1G, today): G=0.067 → P(Contract|LOI) ≈ 0.42×0.067 = 2.8% ; P(Contract) = P_LOI·G·P_raw ≈ 0.43×0.067×0.42 ≈ 1.2%.
• V=0.45 (Dogpatch-brokered vehicle): G=0.30 → P(Contract|LOI) ≈ 12.7% (P(Contract) ≈ 5.4%).
• V=0.75 (MAHI confirmed prime — the realistic resolver here): G=0.50 → P(Contract|LOI) ≈ 21.1% (P(Contract) ≈ 9.1%).
• V=1.00 (Stamp-4, incorporated, T→1.0): G=1.0 → P(Contract|LOI) ≈ 42.2% (P(Contract) ≈ 18.1%).

Score100 = S1·20 = 74.0. Board anchor was [PRIOR-derive] with value = very-high. Reconciliation: 74.0 sits at the top of the European cluster and above the comparable MAHI-maritime row (66.4) — justified because, unlike a domestic buy, this opportunity has a live funded mandate on a clock (mandate_pull=5, funding_to_pay=5) and a non-dilutive route — while the gate (V) remains the binding constraint exactly as it is for every package. Derived band: to-LOI ~38–48%; to-buy ~9% at V=0.75 (MAHI prime) rising to ~18% at V=1.0. No null required — this is a confirmed opportunity with a real, cited procurement route. (MAHI is the gate-resolver here, as in b01; its value lifts the gate, not Stage 1.)

§19 — Commercial

This is a non-dilutive grant play, not a SaaS sale, so the "commercials" are the WP funding share, with the production pricing band noted only for context:

• WP4 funded share (the actual money in scope): indicative €0.8–1.4M [TBD on WP scope] over the project, 100% EU-funded (Research Action) — paid via the consortium, no co-funding required from Nexus. ACV is non-recurring grant, not subscription.
• EUDIS Business Accelerator (parallel): €120k seed voucher. [verified]
• Downstream production pricing (context, if a national buyer later procures the delivered decision-layer): Tier-1 €120–240k/yr (one national CISE/seabed-COP node) → Tier-2 €600k–1.2m/yr (multi-agency) → Tier-3 €2.4–4.8m/yr (EU-scale / multi-MS COP).
• Terms & requirements: EDF Grant Agreement accession via the prime (MAHI); Nexus Synergy must be an incorporated EU beneficiary (or accede via the prime) to draw funds; a consortium agreement with the IP/background-IP carve-outs; FRIA + dual-use export classification before submission.

§20 — Legal blockers

• X1 — Stamp-1G cap (applied): the founder cannot be a company director/shareholder or a signing/invoicing beneficiary until Stamp-4 (~t_resolve 10 mo here). Mitigation specific to this play: MAHI as coordinator/prime carries the consortium signature and lifts the gate (V 0.10→0.75); Nexus accedes as a beneficiary once incorporated (EEA-resident director Patrick/Manuel + €25k s.137 bond, or post-Stamp-4). ⚖️ CONFIRM (immigration solicitor). [verified — irishimmigration.ie]
• X6 — NATO DIANA exclusion (applied): Ireland is non-NATO ⇒ no DIANA route; EU vehicles only (EDF/EUDIS/PESCO) — which is precisely the route chosen here. [verified — diana.nato.int]
• Irish neutrality posture (opportunity-specific): participation in an EDF defence fund is constitutionally sensitive; Nexus' WP is scoped to protective C2 only — a decision/audit overlay, never an effector — the neutrality-compatible framing, recorded in the DoA. Lisbon-Protocol neutrality is a framing constraint, not an eligibility bar (Ireland is an eligible MS). ⚖️ CONFIRM (counsel, for the public posture). [verified — Lisbon neutrality protocol]
• X4 — Dual-use export (Reg (EU) 2021/821): the integrated seabed-monitoring configuration requires a dual-use classification check against the Annex-I control lists, plus screening under the Article 5 cyber-surveillance catch-all — which bites only where an exporter is aware an item is intended for covert surveillance of natural persons or for internal-repression / serious human-rights misuse (per the Commission's 15 Oct 2024 Art-5 Guidelines). Nexus Synergy's protective, track-level-only C2 (person_identifier IS NULL, no covert surveillance of persons) is outside the cyber-surveillance definition by construction, but a per-export classification + end-use due-diligence determination is still required for the assembled configuration. ⚖️ CONFIRM (export-control counsel). [verified — Reg (EU) 2021/821; EC Art-5 Guidelines 15 Oct 2024]
• X2 — EU AI Act: a security-C2 decision-support system is high-risk ⇒ FRIA (Art 27), human-oversight (Art 14), and Art-5 prohibited practices excluded (applied 2 February 2025) by the protective-only verb set + person_identifier IS NULL. [verified]
• X5 — EDF/defence procurement: EDF runs under its own grant rules + Art-12(6) ethics review; Member States may invoke Art-346 TFEU for the most sensitive downstream procurement — anticipate negotiated/exempt procedures for any national follow-on. [verified]
• X7/X8 — pre-incorporation contracting + IP licence chain: sign the consortium commitment letter in the founder's own name (s.45 ratifiable); paper the Adverant→Nexus one-way perpetual licence + the MAHI background-IP carve-out. ⚖️ CONFIRM.

§21 — Warm-intro contact + the specific ask

Contact: Pieter-Jan Note, Co-Founder & CEO, MAHI BV (Belgium) — via the founder's direct founder-to-founder relationship [verified person; email not public]. The specific ask (outreach-drafts §2): "Will MAHI act as coordinator/prime on an all-EU Belgium+Ireland consortium for the EDF-2026 CSIP seabed-infrastructure-protection research action (closes 29 Sep 2026), with MAHI owning WP1 (the USV/sensor edge) and Nexus Synergy owning WP4 (the sovereign COP decision-layer) — and can we book the technical session on the MAHI Sense open API / track schema to scope the WP1↔WP4 data contract?" Secondary asks: a 3rd Member-State sensor/command-centre partner via the EDF Partner Pool, and Patrick Walsh as the Defence-Enterprise (RTI Unit / DEC / IDSA) door on the Irish side. (MAHI's willingness to be prime is a target outcome, not a commitment — confirm in writing before relying on it.)

§22 — Open questions + Sources

Open questions: (1) Will MAHI agree to be coordinator/prime, and on what background-IP terms? — TBD [unverified]. (2) Which 3rd Member-State partner (underwater sensors / command-centre — NL/ES/EE candidates) and via which EDF-Partner-Pool contact? — TBD. (3) The exact CSIP topic conditions (TRL targets, funding-rate split, eligibility of an Irish pre-incorporation entity as beneficiary) — confirm against the full Call-Topic-Descriptions PDF. (4) Dual-use export classification of the integrated configuration — TBD ⚖️ CONFIRM. (5) The public neutrality posture for an Irish entity in an EDF defence consortium — TBD ⚖️ CONFIRM (counsel). (6) Whether to bid the full CSIP topic or a leaner EUDIS SME non-thematic call (RA EDF-2026-LS-DIS-RA-SMERO-NT, €35M) as a faster-to-assemble alternative — TBD.

Sources (dated 2026-06-05):

• EDF 2026 Work Programme adopted 17 Dec 2025; ~€1bn; 10 calls / 31 topics — defence-industry-space.ec.europa.eu/edf-work-programme-2026_en [verified]
• EDF 2026 calls close 29 Sep 2026 17:00 Brussels — hezelburcht.com EDF-2026; zabala.eu EDF-2026 [verified]
• EUDIS €231M; ≈€60M disruptive tech; ≈€60M non-thematic SME; SME research ≤€4M (100%) / development ≤€6M; Business Accelerator €120k voucher — eudis.europa.eu; eudis-business-accelerator.eu [verified]
• EDF/EUDIS consortium minimum: ≥3 legal entities in ≥3 Member States (assoc.: Norway) — eudis.europa.eu/eudis-tracks/sme-calls_en [verified]
• CSIP "Layered Critical Seabed Infrastructure Protection" RA, ~€30M, ≥3 orgs / 3 MS or Norway, integrated command and control + maritime command centres — ketmarket.eu; scientifyresearch.org CSIP topic [verified]
• EMSAS "Enhanced Medium-Size Semi-Autonomous Surface Vessel" DA, ~€90M (context only) — ketmarket.eu EMSAS [verified]
• EUDIS SME non-thematic topic codes (RA EDF-2026-LS-DIS-RA-SMERO-NT €35M; DA EDF-2026-LS-DA-SME-NT €30M) — eudis.europa.eu/eudis-tracks/sme-calls_en [verified]
• Ireland is an eligible EU Member State for EDF; neutrality is a Lisbon-Protocol guarantee (framing constraint, not eligibility bar) — ireland.representation.ec.europa.eu; universitytimes.ie (Mar 2026) [verified]
• Ireland non-NATO ⇒ DIANA excluded ⇒ EU vehicles only — diana.nato.int; legal-blockers-register X6 [verified]
• MAHI BV edge role (MAHI Sense open API, Alpha-8 USV; DSTL/TF59 defence work) — mahi.be; contact-register §B [verified]
• EU AI Act Art-5 prohibited practices applicable 2 February 2025 — AI Act timeline [verified]
• Dual-use export Reg (EU) 2021/821 (dual-use classification + Art-5 cyber-surveillance catch-all; EC Art-5 Guidelines 15 Oct 2024) — legal-blockers-register X4 [verified]
• Stamp-1G bars directorship/self-employment until Stamp-4 — irishimmigration.ie; discovery-playbook open-items [verified]

---