Unwitnessed routes, by contract row — 2026-09-01

Status: work queue. Owner: whichever lane owns the row. Generated from the route register at 3466740e; regenerate by re-running dclutch-route-census inventory and re-deriving status the way tools/genref/generate.mjs does (bindings, then blocked.json).

55 of 161 routes are NEVER-EXECUTED, no stated reason. That number is the largest single thing C-00 and C-16 close against, and it is protocol work rather than bookkeeping: the instrument gap behind it was closed separately (10 rows repaired at 96ddf38f, 2 more witnessed at 939d0806). Every route below has no campaign in the tree at all.

A route sits in the row whose capability it serves, not the program that hosts it, because the lane that would drive it is the capability's lane. The mapping is hand-authored and reviewable — if your row lists a route you do not own, say so and it moves.

Two things this list does NOT say. It does not say these routes are broken: an unwitnessed route is one no campaign drives, which is a statement about coverage and not about correctness. And it does not say they are all equally hard — several are one campaign apart, and blocked.json is the right home for any that turn out to be structurally undrivable, with a reason and an owner, rather than being left in this list to look like unstarted work.

---

C-01 — Infrastructure, Registry, release lineage, migration and Rent (4)

C-02 — Compiler-shaped product entrance (5)

C-04 — Direct (4)

C-06 — Dealer (8)

C-07 — Series (2)

C-08 — Structured/Fractional representation (5)

C-09 — Objective resolution (14)

C-10 — Claims, Custody and terminal lifecycle (13)