🔗 Harness Path — UOR Kappa Mage

16 · uor_kappa_mage — the first descendant — an entry in the origin fleet catalogued in The Harness Paths (group: Descendant lanes). The catalogue is the operator's work WITH the harness, not part of the system you adopt; none of its code lives in the harness repo. This is the map entry, so you can see what a filled config looks like at this weight.

(public repo github.com/mitchuski/uor-kappa-mage — a contribution lane,
not an optimization loop, and it says so on every page)

This entry is a new kind. It carries no harness.config.mjs, no
frontier, no Gap draw, no VALIDATED/MIRAGE ladder — and it is not partial,
because what it took from the skeleton is not the loop but the
constitution, inherited whole. Its fleet method states it plainly: two
documents govern every fleet session, and this lane inherits both —
TRUSTS.md T1–T6 and GROUND_RULES.md GR-1..GR-10 — and its corpus
register lists this repo's HOLONS.md, tools/kappa.mjs, tools/vrc.mjs,
and tools/holon_audit.mjs as PUBLIC source. The lane's work: convergence
and overlap research notes against a live upstream substrate
(UOR-Foundation/kappa-registry, pinned at a named commit), staffed by the
Programme's fourteen role cards (A0–A13; the role that drafts never
verifies its own citations), folding one-way and review-gated. Its first
fleet run seated twelve verifiers — one per convergence section — and all
twelve returned HOLDS_WITH_CORRECTIONS, the corrections filed at
win-prominence (including the one that cut its own framing down: κ's
lineage is from UOR, so the substrate is shared lineage, not parallel
invention; and the upstream's DelegationScope is ahead of the VRC —
a gap filed as prominently as any convergence, GR-6). It even cites this
catalogue back: entry #10's harness-vs-auditor boundary decides its own
next move — citation checks are auditor work; whether a convergence
thesis survives hostile reading is harness work, and the refuter fleet
for that is designed and waiting on the door. The inheritance was proven
where it counts: the lane went public only on the First Person's explicit
word, after two refusals — T6, working in a repo that contains not one line
of the engine.

At a glance —
objective: none to minimise — convergence understood before anything is proposed upstream; understanding as key · gate: the Programme's P-gates + a per-section verification fleet; verdict lexicon HOLDS_WITH_CORRECTIONS, corrections folded inline · hard constraint: the fold is one-way and review-gated — nothing lands upstream except through maintainer review, and the lane changes no upstream code
the Gap: not drawn — inherited: the constitution holds the seats apart, and the planned refuter fleet (designed, unrun) is where the lane's theses will face an adversary · lenses: fourteen Programme roles, A0–A13 — drafter never verifies its own citations · canary: the correction record itself — a first run that returned no corrections would indict the fleet, not flatter the notes
weight: descendant lane · public · the lesson it carries: the skeleton's deepest export is not the loop but the constitution — a lane with no metric, no engine, and no Gap can still inherit every trust, and the proof is behavioural: the door held twice before it opened.

Navigation

← Welcome Visitors · The Harness Paths · Adopting The Harness · Workflows And Runtimes

the wiki is a projection; ~/dual-agent-harness (git) is the source of truth · (⚔️⊥⿻⊥🧙)😊 = neg ⊕ bnot → succ