diff --git a/docs/trajectories/zeta-name-audition/RESUME.md b/docs/trajectories/zeta-name-audition/RESUME.md
new file mode 100644
index 0000000000..fce049f607
--- /dev/null
+++ b/docs/trajectories/zeta-name-audition/RESUME.md
@@ -0,0 +1,92 @@
+# Trajectory — ζ name-audition: from wordmark to a measurement of our own machine
+
+Status: **active**, shadow-lane. Last refreshed: 2026-07-02 (Otto, shadow\*).
+Authorization: Aaron — *"push forward safely with any math even if it seems core;
+it's based on our name, if we get it wrong it's not our identity, we rotate in
+another function."* And *"move both of those forward and save their resume file in
+case we crash — this is very cool."*
+
+Crash-safety note: this file exists BECAUSE Aaron asked to persist the work-front and
+the framings he was streaming, so a session crash loses nothing. Preserve the ferries
+below verbatim in spirit.
+
+## Why this exists (the thesis)
+
+The name "Zeta" was conferred before it had meaning (Kenji proposed ~4 names off the
+early DBSP F# + proofs; Aaron chose Zeta). This trajectory is the name **auditioning
+into a construction**: conferred label → captured entropy → earned identity. Each
+slice makes a real zeta function executable + self-verified, escalating from outside
+objects to **our own machine**.
+
+## Landed slices (all self-verified: two independent computations must agree)
+
+1. **Commutative slice** — #9146 (fleet-Otto, main-lane). Euler product over knots
+   under connected sum (Schubert unique factorization); `Σ N(K)⁻ˢ = Π(1−N(p)⁻ˢ)⁻¹`
+   exact to degree 40; converse locked. `tests/…/ZetaOverPrimeShapes.Tests.fs`.
+2. **Noncommutative slice (Ihara)** — #9148. Ihara zeta of a graph; primes =
+   primitive closed geodesics (compose noncommutatively). Geodesic side
+   `exp(Σ tr(Wᵏ)uᵏ/k)` = Bass side `(1−u²)^{r−1}det(I−Au+Qu²)` to degree 24 over K₄.
+   `tests/…/IharaZeta.Tests.fs` + doc `…ihara-zeta-the-noncommutative-upgrade…md`.
+3. **Dynamical slice (Artin–Mazur) — ζ OF THE SCHEDULER** — #9151. The cell scheduler
+   IS a deterministic dynamical system; `ζ = exp(Σ Fix(fᵏ)uᵏ/k)` = `Π_periodic-orbit
+   1/(1−u^|O|)`. Round-map `M=I+S` over `GF(2)^N` (ring of bits, each cell integrates
+   its left neighbour mod 2 = the synchronous DoP-invariant round). Unique fixed
+   point = the quiescent all-zeros society = ζ's leading 1. Self-verified to degree
+   |S|. `tests/…/SchedulerDynamicalZeta.Tests.fs` + doc `…artin-mazur-zeta-of-the-
+   cell-scheduler-round-map.md`.
+
+## In flight (move-forward #2, not yet landed)
+
+- **Ihara over the braided catalog directly** — replace K₄ with the graph of the
+  catalog's braid generators, so geodesics are real braid words (weights: crossing=1,
+  plait-move=3, braid=6, per #9146). Ties #9148 to #9146's actual objects.
+
+## Ferries from Aaron (2026-07-02 — PRESERVE; these are the vision, not built yet)
+
+- **Periodic orbits = memory-space orbits made visible.** They are exactly what a
+  **cheat engine** finds by pattern-matching **ISR / interrupt-handler cycles**
+  (Arrow-category captured state) — "usually character loops" (a sprite's walk-cycle
+  IS a period-k orbit in memory). The Artin–Mazur zeta is the **formal, math-grounded
+  version of the cheat-engine's orbit detection**: the zeta *enumerates* what the
+  ad-hoc scanner *finds*.
+- **We already do this ad-hoc.** CHIP-8 + `IScheduler` already predict themselves —
+  but ad-hoc, not this formal-math-based. This trajectory is the **rigorous basis for
+  existing capability**, not a new feature. Existing surfaces:
+  `src/Core/Chip8PredictionRoom.fs`, `Chip8Observer.fs`, `PredictionScheduler.fs`,
+  `DarkHallScheduler.fs`.
+- **mod-2 society = James Gates' adinkras.** `GF(2)^N` + mod-2 addition is exactly
+  where Gates' **doubly-even self-dual ECC** (adinkra codes) live. Cell configs can be
+  **secret / generator code words for society members, by purpose** — the codeword IS
+  the member's identity/generator. Ties to rule
+  `only-the-irreducible-is-primitive-generate-the-rest` (the generator IS the ECC;
+  adinkra→Clifford→E8) and `src/Core.Lean4/Lean4/CayleyDicksonDoublyEven.lean`.
+
+## Routed next steps (priority order)
+
+1. **Ihara over the braided catalog** (move-forward #2, committed).
+2. **Formalize the ad-hoc CHIP-8/IScheduler self-prediction**: Artin–Mazur zeta over
+   the actual CHIP-8 VM step-map / a scheduler configuration, making memory-space
+   character-loop orbits formally enumerable — the math-grounded cheat-engine.
+   Connect to `Chip8PredictionRoom` / `PredictionScheduler`.
+3. **Adinkra codewords as member identity**: use the doubly-even self-dual code over
+   `GF(2)^N` to assign purpose-based generator codewords to society members; the
+   dynamical zeta's orbits then classify member behaviour.
+4. **Wire the zeta into the soft `IScheduler`** so the loop can predict its own
+   recurrence spectrum (transient vs. recurrent, orbit periods) before running.
+5. Further math rungs (routed, not rushed): Bartholdi / Ihara–Selberg (2-variable),
+   Ruelle dynamical zeta with weights, Milnor (Alexander = Lefschetz zeta), Kurokawa
+   zeta-of-categories (the braided-monoidal-category seat).
+
+## Discipline
+
+Every slice is **self-verified**: compute the zeta two independent ways (that share
+no code) and require coefficient-by-coefficient agreement — so a wrong construction
+SHOWS (a wrong operator / determinant / map diverges). This is the safety net that
+makes "push into core math" safe: getting it wrong is visible, not silent.
+
+## Anchors
+
+Euler 1737; Riemann 1859; Schubert 1949 (knots); Ihara 1966, Bass 1992, Hashimoto
+1989, Terras 2010 (graph zeta); Artin–Mazur 1965, Bowen–Lanford 1970, Ruelle, Smale
+(dynamical zeta); S. James Gates Jr. (adinkras / doubly-even self-dual ECC);
+Mazur/Morishita (arithmetic topology, knots↔primes).