The Ecosystem

NavCom is one of two applications. This defines the other, what passes between them, and what must not.

Lineage is in lineage.md. The object both are built from is in attestation.md.

Two altitudes of one idea

Both serve the operator who isn't in the field — the guy in the chair. That archetype is NavCom's central innovation at local scale and Starcom's entire premise at global scale.

You navigate by the stars. Starcom is the fixed reference frame; NavCom is crossing real terrain using it.

Starcom is a 3D global cyber command interface for cyber investigations and strategic intelligence — decentralised, open source. It refines raw intel into Intel Reports, and those into Mission Packages. Its operators appear in NavCom's roster as the Investigator and the Cartographer.

The Earth Intelligence Network

Not "two apps that share data." A network where every claim carries its own evidence of how much to trust it.

The idea inherited from Steele is the eight tribes — intelligence fails from structural non-sharing between groups that each hold a real piece, not from insufficient collection. His actual phrase is information-sharing and sense-making: two acts, and almost everyone builds only the first. The tribes exchange conclusions stripped of their evidence chain, which is why the exchange doesn't compound.

An attestation is the technical answer to that, and it is what both applications already are.

NavCom's contribution is the ninth tribe — and it is sharper than "people with local ground truth." It is people who act without institutional authority and must therefore carry their own provenance. Institutions are believed for what they are. The ninth tribe is believed for what it can show.

Starcom refines what is publicly knowable. NavCom holds what is only knowable by being present.

The convergence

Starcom's Finding carries a claim, a confidence, corroboration with a source count and an algorithm version, and a signature. NavCom's directory record carries a claim, a method, an age, a derived confidence, and an author.

Two teams, no shared code, the same object. That is the strongest available evidence that the primitive is real rather than an aesthetic — and it changes what integration is. This is not a negotiation between two schemas. It is one object, aimed at different subjects, and the work is recognising that rather than translating.

Practical consequences:

• Vocabulary converges on Starcom's where it already exists. claim is the payload; attestation is the signed whole
• A Mission Package handed to NavCom is an attestation about a situation, weighed like any other — not a directive
• NavCom answering Query from directory data attaches provenance, which is the same field Starcom's Finding already carries

The handoffs

Two objects cross, in opposite directions.

   STARCOM   raw intel → Intel Report → Mission Package
      │  ▲
      │  │ patterns, aggregated and de-identified   (outbound)
      ▼  │
   NAVCOM   Console (watch)  ·  Field Terminal

Outbound is specified: C35 — field data crosses as patterns, never as operator records. An operator may contribute one incident to a case without surrendering their log.

Inbound is the harder direction, and three existing rules shape it before its contents are known:

1. It must not become a feed. An inbox that accumulates is a feed with a military haircut, and principle 2 does more anti-drift work than every other rule
2. It is the most likely vector to break invariant 1. OSINT-derived material is very likely to describe individuals
3. It has no home on the Field Terminal. Five screens, one hand, dark, cold

Where a Mission Package lands

The Console, and only the Console. It reaches the field as a Query answer or an Assist, or it doesn't reach the field.

No new primitive: the watch is already the human filter between a large knowledge layer and someone standing in the cold. A package is another thing the watch knows.

• No Mission Package UI on the Field Terminal, ever
• No notification to any operator about one
• It becomes directory or playbook entries if it is durable, or it expires with the board if it is current. Never a fourth tier
• Passing it on is a logged watch action like any other

The valve

Filtering happens on arrival, on NavCom's side, regardless of what the sender believes it removed. Starcom states its own residual gap — redaction masks emails and numbers, not names — and an upstream redactor that documents its own incompleteness cannot be the enforcement point for invariant 1. Stated as a requirement; no ingestion exists yet.

What NavCom must never inherit

• Clearance and compartmentation. lore.md already refused these from the fiction, and an intelligence platform is exactly where they re-enter. NavCom accepts classification: UNCLASSIFIED only — a volunteer network has no business holding more, and taking the field while ignoring it would leave a rank system half-built
• Global scope meeting a local privacy rule. Starcom's job is knowing things about the world. NavCom's job includes not knowing things about people
• The analyst's confidence register. An Intel Report is written to be acted on. A NavCom answer at 10pm shows its age and says "call first" when it doesn't know
• Cloud anything. C26 — local inference only

Open

• What a Mission Package contains. Everything above constrains its shape without knowing its payload, deliberately
• Broadcast versus addressed. Today's implementation is plaintext broadcast, which makes Console-only a convention rather than a boundary. If it becomes addressed, the Watchtower pubkey is handed over out of band by a person — bootstrap.spec.md forbids a discovery endpoint, because a list of Watchtowers is a list of where operators are
• entityRef in a public tag. A Finding's d tag exposes it unencrypted. Location-only today, but attribute is deliberately open-ended
• Key rotation. Neither project has a story. A shared Mk1 problem, not an asymmetry