@rig-c/core 2.39.0 → 2.39.1

package/docs/FACE.md

@@ -33,11 +33,11 @@ hold its results.
   breath's `chest` is a **sibling** of the plate it must not scale, and §4.4 is the
   artless-parent pattern the shared-shift split is one instance of
 - The worked example every number below comes from:
-  [`gallery/portrait`](https://github.com/firejune/rigc/tree/main/gallery/portrait),
+  [`gallery/portrait`](https://github.com/rig-c/core/tree/main/gallery/portrait),
   and its measurement half,
-  [`FINDINGS.md`](https://github.com/firejune/rigc/tree/main/gallery/portrait/FINDINGS.md)
+  [`FINDINGS.md`](https://github.com/rig-c/core/tree/main/gallery/portrait/FINDINGS.md)
 - **The same closed forms on the other axis**:
-  [`gallery/nod`](https://github.com/firejune/rigc/tree/main/gallery/nod) is a
+  [`gallery/nod`](https://github.com/rig-c/core/tree/main/gallery/nod) is a
   worked `pitch`, and this page stays written for a **yaw**. It re-derives §4.2's
   fold angle on uneven *rows* and brackets it against `A39` at 33°/34°, and it
   measures §5's foreshortening at **0.863–1.176** against the yaw's 0.892–1.064
@@ -46,7 +46,7 @@ hold its results.
 - **The turn driven by a value instead of played as a time** — the same keys, on
   an axis: §8's *The turn as a value rather than a time* derives the range,
   **AUTHORING §3.5.2** is the `slider` constraint that does it, and
-  [`gallery/look`](https://github.com/firejune/rigc/tree/main/gallery/look) is
+  [`gallery/look`](https://github.com/rig-c/core/tree/main/gallery/look) is
   the worked case. Read it when the angle has to follow something outside the
   animation — a pointer, a gaze target, a game value
 
@@ -402,7 +402,7 @@ authority to say what it measured, and the step share is that.
 
 📐 Method, harness and the full ladders live in the repository rather than in
 this package, as
-[`bench/studies/2026-09-05-density`](https://github.com/firejune/rigc/tree/main/bench/studies/2026-09-05-density) —
+[`bench/studies/2026-09-05-density`](https://github.com/rig-c/core/tree/main/bench/studies/2026-09-05-density) —
 the same study also measures why `contour` is the wrong generator for a turn: it
 saturates at 868 vertices however fine the tolerance, its vertices all sit on the
 silhouette so it samples **2 %** of the depth range, and its ear-clipped interior
@@ -424,7 +424,7 @@ step. ⇒ **Author the sheet so it changes by at least ~8 levels across one mesh
 cell**, and spend its whole 0–255 range on the part — a map using an eighth of
 the range describes the same surface and reports 5.8° less of it. Method and
 ladders:
-[`bench/studies/2026-09-05-noise`](https://github.com/firejune/rigc/tree/main/bench/studies/2026-09-05-noise).
+[`bench/studies/2026-09-05-noise`](https://github.com/rig-c/core/tree/main/bench/studies/2026-09-05-noise).
 
 ⭐ **And you do not have to guess which of the three you are looking at.** The
 lines under the ceiling say it: the **1st percentile over
@@ -736,7 +736,7 @@ practical shape is checkable: **key the fade to 0 at or before
 the last angle that gates green, and let the fold happen after it.**
 
 📘 **The identity above, used forwards on a `pitch`.**
-[`gallery/nod`](https://github.com/firejune/rigc/tree/main/gallery/nod)
+[`gallery/nod`](https://github.com/rig-c/core/tree/main/gallery/nod)
 (repository material) picks its ceiling first and solves the *rows* out of it —
 `|y|outer = R·cos θmax` with θmax chosen at 21° giving 140.037 — and then
 brackets the fold this section predicts against `A39` itself: **33° gates green
@@ -793,7 +793,7 @@ The far eye narrows to 89%, the near eye widens to 106% (**derived**: 0.8922 and
 
 📘 **How much this is worth depends on the axis, and there is a worked case for
 the other one.** The same closed form on the `pitch` of
-[`gallery/nod`](https://github.com/firejune/rigc/tree/main/gallery/nod)
+[`gallery/nod`](https://github.com/rig-c/core/tree/main/gallery/nod)
 (repository material) spans **0.863 … 1.176** across its five features, against
 the **0.892 … 1.064** above — a wider span at the *identical* angle, because
 `α = atan2(coordinate, depth)` grows with the coordinate and a face is taller
@@ -883,7 +883,7 @@ are all timing:
   is a third drawing.
 - **A swap has to dodge the frame grid.** A stepped key exactly on a sample time
   can be missed by a player accumulating `1/fps`, which is why
-  [`gallery/squash`](https://github.com/firejune/rigc/tree/main/gallery/squash)'s
+  [`gallery/squash`](https://github.com/rig-c/core/tree/main/gallery/squash)'s
   blink key sits at `0.399999` (AUTHORING §4.5). **A continuous channel does not
   care where the samples land.**
 
@@ -1112,7 +1112,7 @@ duration   2 x range x scale
 time       to + (degrees - from) x scale    the slider's own mapping
 ```
 
-📐 **Worked, on [`gallery/look`](https://github.com/firejune/rigc/tree/main/gallery/look)**,
+📐 **Worked, on [`gallery/look`](https://github.com/rig-c/core/tree/main/gallery/look)**,
 whose face is a 21 × 9 `grid` over one depth sheet:
 
 ```bash
@@ -1323,7 +1323,7 @@ exactly as they do above. Write `"additive": true` on all of them.
 
 ✅ **The editor half, measured.** The round trip was taken with
 `tools/editor_roundtrip.ts` on a licensed editor (data version 4.3.26) against a 4.3.13 build of
-[`gallery/look`](https://github.com/firejune/rigc/tree/main/gallery/look) — this
+[`gallery/look`](https://github.com/rig-c/core/tree/main/gallery/look) — this
 subsection's worked case, not the `gallery/portrait` this page names at the top,
 which declares no constraints at all and so can carry no slider. What it found:
 
@@ -1784,16 +1784,16 @@ substitute for looking at three scales.
 
 ## 10. The worked example
 
-[`gallery/portrait`](https://github.com/firejune/rigc/tree/main/gallery/portrait)
+[`gallery/portrait`](https://github.com/rig-c/core/tree/main/gallery/portrait)
 is this page on real art — **22 parts, 27 bones, 22 slots, 2 meshes, 40
 vertices**, three animations, and a README that derives every number rather than
 listing it. Its
-[`FINDINGS.md`](https://github.com/firejune/rigc/tree/main/gallery/portrait/FINDINGS.md)
+[`FINDINGS.md`](https://github.com/rig-c/core/tree/main/gallery/portrait/FINDINGS.md)
 is the measurement half: what it cost, the seven-angle sweep, the five tool gaps
 it filed.
 
 📘 **The `pitch` has its own worked case**, and this page deliberately stays a
-yaw: [`gallery/nod`](https://github.com/firejune/rigc/tree/main/gallery/nod) is
+yaw: [`gallery/nod`](https://github.com/rig-c/core/tree/main/gallery/nod) is
 the same two closed forms on the other axis, plus a travelling `wave` (AUTHORING
 §4.11.1). §4.2 and §5 above point at the two places its figures are worth
 reading beside these — the fold angle bracketed against `A39`, and a

package/docs/INGEST.md

@@ -644,7 +644,7 @@ it hands you, and every rule in them applies to a spec `ingest` wrote.
 
 A file from another Spine generation is the one case where transcription is not the
 first fallback: migrate it through the editor of its own generation first, as
-[GENERATIONS.md](https://github.com/firejune/rigc/blob/main/docs/GENERATIONS.md) §4
+[GENERATIONS.md](https://github.com/rig-c/core/blob/main/docs/GENERATIONS.md) §4
 states.
 
 ### 2.1 The workflow
@@ -707,7 +707,7 @@ the second is the one that costs a day:
   all, and then you are debugging your own incomplete work rather than the format.
 
 ⚠️ **When a kind turns out not to be expressible, stop and say so — that is a finding,
-not a blocker to route around.** [SURVEY_2026-08-22.md](https://github.com/firejune/rigc/blob/main/docs/SURVEY_2026-08-22.md) is the
+not a blocker to route around.** [SURVEY_2026-08-22.md](https://github.com/rig-c/core/blob/main/docs/SURVEY_2026-08-22.md) is the
 per-skeleton survey of exactly this. ⇒ Check the survey for your feature before
 concluding either way, and if it is genuinely absent, the shape of the answer is *"this export
 uses X, which the motion spec cannot say"* with a pointer — not a silent
@@ -869,7 +869,7 @@ catches a whole-figure scale error, and it is the only one that does.
 ### 2.4 The worked precedent, and what to take from it
 
 Three transcriptions of official Spine exports live in this repository under
-[`bench/transcriptions/`](https://github.com/firejune/rigc/tree/main/bench/transcriptions/)
+[`bench/transcriptions/`](https://github.com/rig-c/core/tree/main/bench/transcriptions/)
 — `3-timing-and-spacing` (3 bones, 2 slots, 2 animations, 69 keys, every curve raw),
 `6-arcs` (weighted meshes, mesh `edges`, four 4.3 transform constraints) and
 `spineboy` in both `ess` and `pro`. They are **worked examples you are meant to read**,
@@ -1482,7 +1482,7 @@ every directory the fetch produced.
 the per-directory terms.
 
 The transcription these recipes start from is
-[`bench/transcriptions/3-timing-and-spacing/`](https://github.com/firejune/rigc/tree/main/bench/transcriptions/3-timing-and-spacing/),
+[`bench/transcriptions/3-timing-and-spacing/`](https://github.com/rig-c/core/tree/main/bench/transcriptions/3-timing-and-spacing/),
 unmodified. The re-pivoted, renamed and extended variants in §4 were built from copies
 of it; none is committed, because each is an illustration of an edit rather than a
 transcription of anything.

package/docs/MOTION.md

@@ -105,7 +105,7 @@ and each fails in its own way rather than merely reading oddly:**
 ⭐ **What is still this page's job is whatever moves the dial**, and that is an
 ordinary animation with §3 applying to it unchanged. The split is visible in the
 worked example: in
-[`gallery/look`](https://github.com/firejune/rigc/tree/main/gallery/look) the two
+[`gallery/look`](https://github.com/rig-c/core/tree/main/gallery/look) the two
 lookup tables `turn` and `tilt` carry **no easing at all**, and `sweep` — the one
 animation there meant to be played — carries every `ease` in the file.
 
@@ -751,7 +751,7 @@ What *is* real: every command line below was run, and every figure printed in an
 output block is what the command actually printed.
 
 🖼️ **For the same recipe on art that ships, the
-[`gallery/`](https://github.com/firejune/rigc/tree/main/gallery) examples are worked
+[`gallery/`](https://github.com/rig-c/core/tree/main/gallery) examples are worked
 in-betweening material** — `walk` is §3.5's arcs and §3.7's phase offsets on two leg
 chains, `ride` puts the same offsets in `groups` + `stagger`, and `squash` is §3.9's
 pivot written as a `deform` about a contact point. Each README says what every key is

package/docs/PROMPTING.md

@@ -8,9 +8,9 @@ measured case: in August 2026 a deliberately small model — Gemini 3.7 Flash on
 the Antigravity harness — authored a benchmark rung over a **3 h 01 m** session,
 observed by the operator; the harness kept no tool transcript, so the run is
 recorded **Unverified** on honesty, with two of its four tiers *unevidenced*
-([the verdict](https://github.com/firejune/rigc/blob/main/docs/pilot/verdicts/2026-08-24-rung3-1.md)
+([the verdict](https://github.com/rig-c/core/blob/main/docs/pilot/verdicts/2026-08-24-rung3-1.md)
 states all of that). Every stumble below is one it actually made, priced by the instruments in
-[PILOT.md](https://github.com/firejune/rigc/blob/main/docs/PILOT.md)'s protocol. A large model makes fewer of these mistakes
+[PILOT.md](https://github.com/rig-c/core/blob/main/docs/PILOT.md)'s protocol. A large model makes fewer of these mistakes
 unprompted; the clauses cost you three lines and make the small model's run
 land and the large model's run land cleaner.
 
@@ -27,7 +27,7 @@ For ordinary use (not a benchmark), the whole prompt is short:
 
 Then add the clauses below. If you are running the **benchmark**, do not
 improvise a prompt at all — hand the agent
-[pilot/rung3-runner-prompt.md](https://github.com/firejune/rigc/blob/main/docs/pilot/rung3-runner-prompt.md) verbatim; it
+[pilot/rung3-runner-prompt.md](https://github.com/rig-c/core/blob/main/docs/pilot/rung3-runner-prompt.md) verbatim; it
 carries the honesty rule, and an improvised summary of that rule is how a run
 stops being scorable.
 
@@ -101,7 +101,7 @@ prompt for the six clauses above and then let the agent work.
 
 ## If you are scoring the run
 
-That is a different document: [PILOT.md](https://github.com/firejune/rigc/blob/main/docs/PILOT.md) is the protocol,
+That is a different document: [PILOT.md](https://github.com/rig-c/core/blob/main/docs/PILOT.md) is the protocol,
 `pilot/rung3-runner-prompt.md` is the exact prompt, and the evaluator sheet it
 names is for the judge only. The one rule that reaches this page: **never put
 baseline figures, reference structure, or previous runs' measures into the

package/docs/RIGGING.md

@@ -105,7 +105,7 @@ beside the offsets it exists to make visible:
 > *"Every limb attachment is OFFSET from its bone, and that is what makes the hinge
 > visible: art centred on its own pivot turns in place, so a search over one angle
 > would move nothing and a wrong hinge would cost nothing either."*
-> — [`selftest.ts`](https://github.com/firejune/rigc/blob/main/selftest.ts), the chain-fit fixture
+> — [`selftest.ts`](https://github.com/rig-c/core/blob/main/selftest.ts), the chain-fit fixture
 
 ⇒ **A correctly placed pivot also collapses the spec.** A pendulum whose bone sits
 on its hinge needs one `rotate` track and no translate at all; the same part with
@@ -144,7 +144,7 @@ never:
   it**, so the bone points at its own mass — at the cost of two emitted fields.
 
 > *"Neither is checkable from the frames."* — rung 3's first attempt
-> ([`2026-08-23-rung3-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-23-rung3-1/LOOP.md) §11)
+> ([`2026-08-23-rung3-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-23-rung3-1/LOOP.md) §11)
 
 ⇒ Pick one, and say in your log that you picked it. AUTHORING §10.1's naming
 convention is the same shape of decision and is worth far more (it decides five of
@@ -162,7 +162,7 @@ rigid.
 > bone you scale-key must sit at that bone's origin, or the offset rides the
 > scale."*
 > — spineboy attempt 4
-> ([`2026-08-28-spineboy-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-28-spineboy-1/LOOP.md) §8)
+> ([`2026-08-28-spineboy-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-28-spineboy-1/LOOP.md) §8)
 
 ⇒ This is the one exception to §1's rule, and it is not a contradiction: a bone
 whose **scale** is the animated property is not a hinge, so there is no hinge to
@@ -184,10 +184,10 @@ None of them is an error message. Each one is a number that reads as ordinary.
 
 | Signature | What it actually is | Record |
 | --- | --- | --- |
-| **A gap that letting the parts off the hierarchy closes.** `idle` fitted to 10.3 union MAE on the frame the setup pose came from and **23.9** eight frames later; a short relax off the hierarchy recovered it to **14.6** | *"A gap that a relax closes is not a pose the rig cannot reach; it is a pivot in the wrong place."* | spineboy attempt 1 ([`2026-08-23-spineboy-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-23-spineboy-1/LOOP.md) §8.2) |
-| **A fitted centre that drifts monotonically with the fitted scale.** The per-frame fits wanted each shadow's centre 8–12 units lower whenever the shadow was small | a part scaling about a pivot below its own centre — *"a pivot you have not modelled, not motion you have measured"* | rung 1 second attempt ([`2026-08-26-rung1-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-26-rung1-1/LOOP.md) §2.3) |
-| **A knob resting exactly on its bound, on one orientation only.** `torso.x` pinned at −35.0 against a ±35 box on the lying frames and nowhere else | the optimum is outside the box, because a pivot error `e` needs a compensation `−R(θ)·e` that grows with the local angle — *"invisible upright, saturating in the lying poses"* | spineboy attempt 5 ([`2026-08-28-spineboy-2`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-28-spineboy-2/LOOP.md) §3) |
-| **An exact fit with no angular spread.** Two ankle joints solved at **rms 0.00 px** with a relative-angle spread of **0° and 2°** | a perfect fit to a system with no unique answer | spineboy, 2026-09-03 first attempt ([`2026-09-03-spineboy-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-09-03-spineboy-1/LOOP.md) §3.3) |
+| **A gap that letting the parts off the hierarchy closes.** `idle` fitted to 10.3 union MAE on the frame the setup pose came from and **23.9** eight frames later; a short relax off the hierarchy recovered it to **14.6** | *"A gap that a relax closes is not a pose the rig cannot reach; it is a pivot in the wrong place."* | spineboy attempt 1 ([`2026-08-23-spineboy-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-23-spineboy-1/LOOP.md) §8.2) |
+| **A fitted centre that drifts monotonically with the fitted scale.** The per-frame fits wanted each shadow's centre 8–12 units lower whenever the shadow was small | a part scaling about a pivot below its own centre — *"a pivot you have not modelled, not motion you have measured"* | rung 1 second attempt ([`2026-08-26-rung1-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-26-rung1-1/LOOP.md) §2.3) |
+| **A knob resting exactly on its bound, on one orientation only.** `torso.x` pinned at −35.0 against a ±35 box on the lying frames and nowhere else | the optimum is outside the box, because a pivot error `e` needs a compensation `−R(θ)·e` that grows with the local angle — *"invisible upright, saturating in the lying poses"* | spineboy attempt 5 ([`2026-08-28-spineboy-2`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-28-spineboy-2/LOOP.md) §3) |
+| **An exact fit with no angular spread.** Two ankle joints solved at **rms 0.00 px** with a relative-angle spread of **0° and 2°** | a perfect fit to a system with no unique answer | spineboy, 2026-09-03 first attempt ([`2026-09-03-spineboy-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-09-03-spineboy-1/LOOP.md) §3.3) |
 
 🚨 **The last one is the trap in its purest form: a zero residual reads as success
 in every other context on this toolchain.** Here it means the estimator was handed
@@ -219,7 +219,7 @@ across all eight shots:
 > 21-unit-different answer that fits the upright frames just as well. The joint is
 > ill-conditioned without the lying poses, which is how the inherited triangulation
 > (idle zoom-read) went wrong without any frame saying so."*
-> — [`2026-08-28-spineboy-2`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-28-spineboy-2/LOOP.md) §5
+> — [`2026-08-28-spineboy-2`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-28-spineboy-2/LOOP.md) §5
 
 Two answers 21 art units apart, at residuals that do not separate them. That is
 what *ill-conditioned* means with a picture attached.
@@ -296,7 +296,7 @@ found that **every one of them measured a basin at the sweep's own cap** — ±4
 > *"On this figure the joint offsets are **weakly identified at the scale a frame
 > can see**, and the honest reading is that they are a prior with a measured width
 > rather than a measurement."*
-> — [`2026-09-03-spineboy-2`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-09-03-spineboy-2/LOOP.md) §3.5
+> — [`2026-09-03-spineboy-2`](https://github.com/rig-c/core/blob/main/bench/runs/2026-09-03-spineboy-2/LOOP.md) §3.5
 
 ⇒ That sentence is the deliverable when the pixels are silent. It is not a hedge:
 a width is a number, and the next attempt can tell whether its own change is inside
@@ -463,7 +463,7 @@ The failure, watched happening:
 > rear-thigh −199.8°`, whose net world rotations are −3°, +31° and −19°: the picture
 > was right and the numbers were unusable. A rotate series that swings 180° between
 > two keys does not interpolate, it spins the figure through a whole turn."*
-> — spineboy attempt 1 ([`2026-08-23-spineboy-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-23-spineboy-1/LOOP.md) §8.4)
+> — spineboy attempt 1 ([`2026-08-23-spineboy-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-23-spineboy-1/LOOP.md) §8.4)
 
 ### 4.1 ⚠️ The exact fold has a precondition, and it is easy to miss
 
@@ -474,7 +474,7 @@ AUTHORING §10.3 prescribes is not a null operation — it costs pixels:
 > *"the hip's three children sit 9–13 units off it, so turning the hip moves their
 > origins and the render. Folding cost MAE on every frame it touched (`idle` mean
 > 23.0 with the fold against 19.9 without, same search)."*
-> — spineboy attempt 2 ([`2026-08-23-spineboy-2`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-23-spineboy-2/LOOP.md) §5)
+> — spineboy attempt 2 ([`2026-08-23-spineboy-2`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-23-spineboy-2/LOOP.md) §5)
 
 ⇒ **Test the precondition before folding.** Children at the origin ⇒ exact gauge,
 fold it (the L1 optimum is the *median* of the children's deltas, which is one line).
@@ -521,10 +521,10 @@ that it is *keyed for a reason the arithmetic states*, not left free for a solve
 
 | Bone | Carries | Its children then key |
 | --- | --- | --- |
-| `faceshift` ([`gallery/portrait`](https://github.com/firejune/rigc/tree/main/gallery/portrait/)) | `−R·sin t`, the shift every feature shares in a yaw | only their residual — FACE §3 |
+| `faceshift` ([`gallery/portrait`](https://github.com/rig-c/core/tree/main/gallery/portrait/)) | `−R·sin t`, the shift every feature shares in a yaw | only their residual — FACE §3 |
 | `comet` (rung 6, rung 8) | the whole travel of a ball and its trail | the squash, and the trail's bend |
 | `body` (rung 7, second attempt) | **translation only** — *"it holds no attachment, so keying its rotation as well would be a gauge"* | their own rotations |
-| `ground` ([`gallery/walk`](https://github.com/firejune/rigc/tree/main/gallery/walk/)) | the contact line | the foot targets — §9.1 |
+| `ground` ([`gallery/walk`](https://github.com/rig-c/core/tree/main/gallery/walk/)) | the contact line | the foot targets — §9.1 |
 
 ⭐ **FACE §3's argument for it is an auditing one before it is a rigging one**: a
 residual is 1–6 units where a total is 30–40, *"nobody can eyeball an error in the
@@ -548,13 +548,13 @@ invisible in the numbers.
 > and takes a `translatey` (0 → 3.4). Parent the head chain to the bone that scales
 > and the whole face inflates 1.4% every breath — visible, and invisible in the
 > spec."*
-> — [`gallery/portrait/README.md`](https://github.com/firejune/rigc/blob/main/gallery/portrait/README.md)
+> — [`gallery/portrait/README.md`](https://github.com/rig-c/core/blob/main/gallery/portrait/README.md)
 
 The same decision, on a ball and its trail:
 
 > *"`ball` is a **sibling** of `tail0` under `comet`, not its parent, so squashing
 > the ball does not squash the trail."*
-> — rung 6 ([`2026-08-23-rung6-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-23-rung6-1/README.md))
+> — rung 6 ([`2026-08-23-rung6-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-23-rung6-1/README.md))
 
 ⇒ **The construction is always the same three bones**: an artless parent that
 carries what both share (§4.4), and two children — one that takes the scale and one
@@ -565,7 +565,7 @@ that takes the chain.
 > *"the neck plate has to take a *fraction* of the head's shift, so it cannot be an
 > ancestor of the head. `neckbase` is the shared parent; `neck` carries the plate and
 > `headroll` carries the head"*
-> — [`gallery/portrait/README.md`](https://github.com/firejune/rigc/blob/main/gallery/portrait/README.md), the part table
+> — [`gallery/portrait/README.md`](https://github.com/rig-c/core/blob/main/gallery/portrait/README.md), the part table
 
 ⭐ **And the pivot goes one link up, which is the same edit for a second reason.**
 The same rig keys `headroll` and never `head`: *"a head rotates about the top of the
@@ -638,7 +638,7 @@ The record that named it:
 > setup render is unchanged. walk 0.176 → 0.150, run 0.232 → 0.201, aim 0.212 →
 > 0.186 after refit."*
 > — spineboy attempt 4
-> ([`2026-08-28-spineboy-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-28-spineboy-1/LOOP.md) §6)
+> ([`2026-08-28-spineboy-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-28-spineboy-1/LOOP.md) §6)
 
 ⭐ **The tie-break is on the frames' side.** *"the shot's own extremes are a
 measurement, while segment lengths taken off a folded pose are an estimate — so when
@@ -659,7 +659,7 @@ fits converge, every residual is ordinary, and the figure splays.**
 > leg overshoots the floor by about 90 units and the fitter has to bend it sideways
 > to land … Re-seeding the leg joints inset from the art's ends — thigh 86 units,
 > shin 94 — put the ankle at world y **53.5** against the boot's own 54."*
-> — [`2026-09-03-spineboy-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-09-03-spineboy-1/LOOP.md) §4.4
+> — [`2026-09-03-spineboy-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-09-03-spineboy-1/LOOP.md) §4.4
 
 ⭐ **Why the art misled it, and this generalises:** the art's outer ends are not the
 joints. A rounded cap's centroid sits at the plate's *tip*, while the joint two
@@ -686,7 +686,7 @@ frames put the standing leg at about 233 — a 79-unit excess with a plausible c
 > better rig, and the 79 units are a real knee bend in the reference's own stance
 > rather than an error in this rig. The surgery §8.1 warns about … was therefore not
 > performed, and the per-frame fitting budget was not spent twice."*
-> — [`2026-09-03-spineboy-2/evidence/limb-reach.txt`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-09-03-spineboy-2/evidence/limb-reach.txt)
+> — [`2026-09-03-spineboy-2/evidence/limb-reach.txt`](https://github.com/rig-c/core/blob/main/bench/runs/2026-09-03-spineboy-2/evidence/limb-reach.txt)
 
 ⇒ **The reach check tells you a chain and a shot disagree. It does not tell you
 which one is wrong.** A bent limb and a long limb produce the same excess. One
@@ -736,7 +736,7 @@ form: **is each frame's silhouette an affine image of the part's own drawing?**
 > freedom the mesh was built with — a four-bone chain, weights blended by height
 > only."*
 > — rung 7, first attempt
-> ([`2026-08-28-rung7-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-28-rung7-1/LOOP.md) §5)
+> ([`2026-08-28-rung7-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-28-rung7-1/LOOP.md) §5)
 
 ⇒ Two controls are what make that quotable and they are worth copying: the art
 through a **known affine** read 0.0044 → 0.0007 (the floor), and the art plus a
@@ -772,10 +772,10 @@ not fail — it degenerates:
 > *"The first candidate lifted the foot 56 units on a 130-unit leg: the chain has to
 > fold to a 70-unit span, and a two-bone solve does that by throwing the knee
 > sideways. 28 units — a fifth of the leg — is the version that reads as a step."*
-> — [`gallery/walk/README.md`](https://github.com/firejune/rigc/blob/main/gallery/walk/README.md)
+> — [`gallery/walk/README.md`](https://github.com/rig-c/core/blob/main/gallery/walk/README.md)
 
 The same arithmetic on the other side of the chain, from
-[`gallery/flex`](https://github.com/firejune/rigc/tree/main/gallery/flex/): rotating a rigid panel about a seam opens a wedge
+[`gallery/flex`](https://github.com/rig-c/core/tree/main/gallery/flex/): rotating a rigid panel about a seam opens a wedge
 of `halfHeight × tan(angle)` — about 22 px at 20° on that cloth — so *"the bend
 angles in `motion.json` are chosen against that overlap, not the other way round."*
 
@@ -801,7 +801,7 @@ differ by exactly the bone's own setup `x`/`y`:
 > *"the first spec moved the figure by the setup offset a second time. Rotation and
 > scale needed no correction: their setups are 0 and 1."*
 > — rung 7, second attempt
-> ([`2026-08-28-rung7-2`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-28-rung7-2/LOOP.md) §5)
+> ([`2026-08-28-rung7-2`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-28-rung7-2/LOOP.md) §5)
 
 The same confusion one attempt earlier, from the other end: *"The first fitter used
 0 as every knob's neutral value. `bone.pose.x` is the bone's whole local
@@ -817,7 +817,7 @@ the diff names it in one line."*
 
 ⚠️ **A bone under a rotated parent is not at the parent's position plus its own
 numbers, and a leg chain is exactly that case.** That sentence is a comment in
-[`gallery/stage.ts`](https://github.com/firejune/rigc/blob/main/gallery/stage.ts), which exists because a drawing script
+[`gallery/stage.ts`](https://github.com/rig-c/core/blob/main/gallery/stage.ts), which exists because a drawing script
 needed setup-pose world positions and *"writing them a second time in the drawing
 script is the defect that produces a shadow under nobody's feet, and it is
 invisible in both files — each is internally consistent."*
@@ -993,7 +993,7 @@ designed around it:
 > the chest and the legs as siblings. Chosen for identifiability, not anatomy …
 > Under a pelvis the legs sit above the only reliable anchor and every leg comes back
 > `no-anchor` on every frame — measured, not predicted."*
-> — [`2026-09-03-spineboy-2`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-09-03-spineboy-2/README.md)
+> — [`2026-09-03-spineboy-2`](https://github.com/rig-c/core/blob/main/bench/runs/2026-09-03-spineboy-2/README.md)
 
 That is the only place in the corpus where a tree was chosen for what could be
 measured through it rather than for anatomy or for the art, and it is worth knowing
@@ -1014,7 +1014,7 @@ backwards:
 `bendPositive` combinations were indistinguishable at 4× zoom on the frame where the
 difference is largest. `shin_f.worldX` separated them immediately. Look at frames
 for whether it reads; read bone positions for what it is doing."*
-([`gallery/walk/README.md`](https://github.com/firejune/rigc/blob/main/gallery/walk/README.md))
+([`gallery/walk/README.md`](https://github.com/rig-c/core/blob/main/gallery/walk/README.md))
 
 ### 8.4 ⚠️ And `pose` can be confident about the wrong twin
 
@@ -1041,7 +1041,7 @@ list of structural rules and a list of refusals.
 > at the contact line; `foot_f` and `foot_b` are children of `ground`. Parent a foot
 > target to `hip` and the hip carries its own feet up with it, **the solver reports
 > success, and the figure hovers.**"*
-> — [`gallery/walk/README.md`](https://github.com/firejune/rigc/blob/main/gallery/walk/README.md)
+> — [`gallery/walk/README.md`](https://github.com/rig-c/core/blob/main/gallery/walk/README.md)
 
 Measured on that build, over the stance where `foot_f` is planted: the hip travels
 **7.9 units up and 2.9 sideways** while the foot reads `382.0, 68.0` on every
@@ -1067,15 +1067,15 @@ reaches the target. So:
 ⭐ **This is where the art and the hierarchy stop being two decisions.** Both plates
 have to be **drawn from their own joint**, and the joint's coordinates *inside each
 drawing* are what every `length` and every attachment offset is measured from —
-[`gallery/rigby.ts`](https://github.com/firejune/rigc/blob/main/gallery/rigby.ts) names all four points for one leg, and
-[`gallery/walk/rig.json`](https://github.com/firejune/rigc/blob/main/gallery/walk/rig.json) states them.
+[`gallery/rigby.ts`](https://github.com/rig-c/core/blob/main/gallery/rigby.ts) names all four points for one leg, and
+[`gallery/walk/rig.json`](https://github.com/rig-c/core/blob/main/gallery/walk/rig.json) states them.
 
 ### 9.3 A constraint carries the whole subtree, and that is the point
 
 > *"`cart` is the constrained bone; `wheel_b`/`wheel_f` and `hip` are its children,
 > so the constraint carries the wheels and the whole figure with it and the tangent
 > rotation tilts all three."*
-> — [`gallery/ride/README.md`](https://github.com/firejune/rigc/blob/main/gallery/ride/README.md)
+> — [`gallery/ride/README.md`](https://github.com/rig-c/core/blob/main/gallery/ride/README.md)
 
 📌 Two placement conveniences from the same paragraph, both of the form *put the
 bone where the numbers become trivial*: `root` is at the world origin, *"which is why
@@ -1104,7 +1104,7 @@ default failure mode**, because it drives bones rather than drawing anything.
 > `true`, so any IK timeline that did not restate it overwrote the constraint's value
 > for the whole animation, with the field still in the file and the gate green either
 > way."*
-> — [`gallery/walk/README.md`](https://github.com/firejune/rigc/blob/main/gallery/walk/README.md)
+> — [`gallery/walk/README.md`](https://github.com/rig-c/core/blob/main/gallery/walk/README.md)
 > (rigc stamps the rig's value onto every emitted ik key)
 
 ⇒ **A structural fact stated in the rig can be overwritten by a per-key default in
@@ -1121,7 +1121,7 @@ all**, unanimously, and paid for it in a measure:
 > point at if asked why there are four rather than one … that measure is the price of
 > the rule, and it is worth paying rather than winning by guessing."*
 > — rung 8, first attempt
-> ([`2026-08-23-rung8-1`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-23-rung8-1/LOOP.md) §13)
+> ([`2026-08-23-rung8-1`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-23-rung8-1/LOOP.md) §13)
 
 ⭐ **The cleanest demonstration that the rule costs something real and is still the
 right rule** is in that same run, which authored two rigs against two references:
@@ -1171,7 +1171,7 @@ the draw order"* — which is what lets one part sit in front of its own parent:
 > arrangement that satisfies both things the frames measure: the gun reads unoccluded
 > in `idle`, and the near leg covers it in `walk`."*
 > — spineboy attempt 2
-> ([`2026-08-23-spineboy-2`](https://github.com/firejune/rigc/blob/main/bench/runs/2026-08-23-spineboy-2/README.md))
+> ([`2026-08-23-spineboy-2`](https://github.com/rig-c/core/blob/main/bench/runs/2026-08-23-spineboy-2/README.md))
 
 ⚖️ **But parentage is a real question with a real answer, and the evidence for it is
 smoothness.** Rung 4 modelled a chain's top as a *sibling* of the disc it hangs

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@rig-c/core",
-  "version": "2.39.0",
+  "version": "2.39.1",
   "description": "AI-authored 2D rigging and animation, verified before it is written. rigc compiles a rig spec into skeleton data in the Spine 4.3 format, gates it with named assertions held to a spine-core round trip's verdicts, and emits nothing that fails. Ships as an agent skill.",
   "type": "module",
   "bin": {
@@ -37,10 +37,10 @@
   "author": "firejune",
   "repository": {
     "type": "git",
-    "url": "https://github.com/firejune/rigc.git"
+    "url": "https://github.com/rig-c/core.git"
   },
-  "homepage": "https://github.com/firejune/rigc#readme",
-  "bugs": "https://github.com/firejune/rigc/issues",
+  "homepage": "https://github.com/rig-c/core#readme",
+  "bugs": "https://github.com/rig-c/core/issues",
   "keywords": [
     "spine",
     "spine2d",

package/skills/rigc/SKILL.md

@@ -109,25 +109,25 @@ model space, its UVs, its triangles — which is a mesh attachment with those th
 (AUTHORING §3.4, *Mesh attachment*); a region at the drawable's bounding-box centre
 keeps only the box. rigc reads none of those formats — FACE §11, the paragraph
 that opens *No Live2D file is read or written*
-([FACE.md](https://github.com/firejune/rigc/blob/main/docs/FACE.md#11-non-goals--stated-so-nobody-proposes-them-as-gaps)) —
+([FACE.md](https://github.com/rig-c/core/blob/main/docs/FACE.md#11-non-goals--stated-so-nobody-proposes-them-as-gaps)) —
 only the pictures they produce. The rule, the background those frames need and
 what a wrong one costs: AUTHORING §0, *When the source is a foreign player*.
 
 ## Which guide, for which need
 
-Read [AUTHORING.md](https://github.com/firejune/rigc/blob/main/docs/AUTHORING.md) first, whatever the need: the two
+Read [AUTHORING.md](https://github.com/rig-c/core/blob/main/docs/AUTHORING.md) first, whatever the need: the two
 spec files field by field, the emission rules, the loop, and the failure map. Then:
 
 | The request is… | Open | Skill |
 | --- | --- | --- |
-| a **skeleton** — how many bones, where each pivot sits, what hangs off what | [RIGGING.md](https://github.com/firejune/rigc/blob/main/docs/RIGGING.md) | `rigc-rigging` |
-| a **movement** — an idle, a loop, from this pose to that one | [MOTION.md](https://github.com/firejune/rigc/blob/main/docs/MOTION.md) | `rigc-motion` |
-| a **face** — a blink, a gaze, a head turn a few degrees off axis | [FACE.md](https://github.com/firejune/rigc/blob/main/docs/FACE.md) | `rigc-face` |
-| a **skeleton.json somebody else authored** — read it, repair it, extend it | [INGEST.md](https://github.com/firejune/rigc/blob/main/docs/INGEST.md) | `rigc-ingest` |
-| you are the **person operating** the agent rather than the agent | [PROMPTING.md](https://github.com/firejune/rigc/blob/main/docs/PROMPTING.md) | — |
+| a **skeleton** — how many bones, where each pivot sits, what hangs off what | [RIGGING.md](https://github.com/rig-c/core/blob/main/docs/RIGGING.md) | `rigc-rigging` |
+| a **movement** — an idle, a loop, from this pose to that one | [MOTION.md](https://github.com/rig-c/core/blob/main/docs/MOTION.md) | `rigc-motion` |
+| a **face** — a blink, a gaze, a head turn a few degrees off axis | [FACE.md](https://github.com/rig-c/core/blob/main/docs/FACE.md) | `rigc-face` |
+| a **skeleton.json somebody else authored** — read it, repair it, extend it | [INGEST.md](https://github.com/rig-c/core/blob/main/docs/INGEST.md) | `rigc-ingest` |
+| you are the **person operating** the agent rather than the agent | [PROMPTING.md](https://github.com/rig-c/core/blob/main/docs/PROMPTING.md) | — |
 
 Every guide linked here is in the installed package at `node_modules/rig-c/docs/`,
 which is the copy that matches the rigc you run; the links go to the repository's
 `main`. Inside the Claude Code plugin the same files are at `${CLAUDE_PLUGIN_ROOT}/docs/`.
 Formats, the CLI reference and the licence chain:
-[README.md](https://github.com/firejune/rigc/blob/main/README.md), installed at `node_modules/rig-c/README.md`.
+[README.md](https://github.com/rig-c/core/blob/main/README.md), installed at `node_modules/rig-c/README.md`.

package/skills/rigc-face/SKILL.md

@@ -9,7 +9,7 @@ compatibility: Requires Bun 1.2 or later and the npm package rig-c.
 
 Load this when the request is a **head rather than a body**: a drawn face that
 breathes, blinks, moves its eyes and turns a few degrees off axis. Every rule below
-is owned by [FACE.md](https://github.com/firejune/rigc/blob/main/docs/FACE.md); this file says when to open it and what
+is owned by [FACE.md](https://github.com/rig-c/core/blob/main/docs/FACE.md); this file says when to open it and what
 it will not do for you.
 
 ## Non-negotiables
@@ -40,16 +40,16 @@ differential audit and the three limits it does not lift.
 
 ## Read, in this order
 
-1. [AUTHORING.md](https://github.com/firejune/rigc/blob/main/docs/AUTHORING.md) — the `deform` timeline field by field
+1. [AUTHORING.md](https://github.com/rig-c/core/blob/main/docs/AUTHORING.md) — the `deform` timeline field by field
    and what rigc refuses in it (§4), the failure map (§5–§6), the editor's
    conventions (§10).
-2. [MOTION.md](https://github.com/firejune/rigc/blob/main/docs/MOTION.md) — timing, easing and the offset table (§3),
+2. [MOTION.md](https://github.com/rig-c/core/blob/main/docs/MOTION.md) — timing, easing and the offset table (§3),
    candidates and the ballot (§4–§5). A blink and a gaze are ordinary motion work.
-3. [FACE.md](https://github.com/firejune/rigc/blob/main/docs/FACE.md) — the face's own geometry: the closed form every
+3. [FACE.md](https://github.com/rig-c/core/blob/main/docs/FACE.md) — the face's own geometry: the closed form every
    number in a turn comes from (§1), the hierarchy underneath (§3), what
    foreshortens (§5), and the deform audit gap (§9).
-4. Then [RIGGING.md](https://github.com/firejune/rigc/blob/main/docs/RIGGING.md) for the hierarchy as a general rule
-   rather than this closed form, and [INGEST.md](https://github.com/firejune/rigc/blob/main/docs/INGEST.md) if the head
+4. Then [RIGGING.md](https://github.com/rig-c/core/blob/main/docs/RIGGING.md) for the hierarchy as a general rule
+   rather than this closed form, and [INGEST.md](https://github.com/rig-c/core/blob/main/docs/INGEST.md) if the head
    arrived as a compiled skeleton.
 
 Every guide linked here is in the installed package at `node_modules/rig-c/docs/`,

package/skills/rigc-ingest/SKILL.md

@@ -10,7 +10,7 @@ compatibility: Requires Bun 1.2 or later and the npm package rig-c.
 Load this when what you were handed is **already a skeleton**: a `skeleton.json`
 with its `.atlas` and page images, and a request to understand it, answer a
 complaint about it, re-express it, or extend it. Every rule below is owned by
-[INGEST.md](https://github.com/firejune/rigc/blob/main/docs/INGEST.md); this file says when to open it and what it will
+[INGEST.md](https://github.com/rig-c/core/blob/main/docs/INGEST.md); this file says when to open it and what it will
 not do for you.
 
 ## Non-negotiables
@@ -61,14 +61,14 @@ thing to reach for; transcription by hand is what you fall back on for a constru
 
 ## Read, in this order
 
-1. [INGEST.md](https://github.com/firejune/rigc/blob/main/docs/INGEST.md) — what every command will and will not do with
+1. [INGEST.md](https://github.com/rig-c/core/blob/main/docs/INGEST.md) — what every command will and will not do with
    a foreign file (§0), `ingest` and transcription (§2), what each validator
    complaint means on an export (§3), and the re-pivot, rename and extend
    recipes (§4).
-2. [AUTHORING.md](https://github.com/firejune/rigc/blob/main/docs/AUTHORING.md) — the two spec files the transcription
+2. [AUTHORING.md](https://github.com/rig-c/core/blob/main/docs/AUTHORING.md) — the two spec files the transcription
    targets (§3–§4), the failure map (§5–§6), and the coordinate contract (§11.2).
-3. Then [RIGGING.md](https://github.com/firejune/rigc/blob/main/docs/RIGGING.md) for why the re-pivot edit has the shape
-   it has, and [MOTION.md](https://github.com/firejune/rigc/blob/main/docs/MOTION.md) for the animation you are adding.
+3. Then [RIGGING.md](https://github.com/rig-c/core/blob/main/docs/RIGGING.md) for why the re-pivot edit has the shape
+   it has, and [MOTION.md](https://github.com/rig-c/core/blob/main/docs/MOTION.md) for the animation you are adding.
 
 Every guide linked here is in the installed package at `node_modules/rig-c/docs/`,
 which is the copy that matches the rigc you run; the links go to the repository's

package/skills/rigc-motion/SKILL.md

@@ -10,7 +10,7 @@ compatibility: Requires Bun 1.2 or later and the npm package rig-c.
 Load this when the request is a **movement rather than a skeleton**: a sentence of
 intent, between zero and N pictures of what the movement passes through, and a
 Spine animation somebody would choose coming back. Every rule below is owned by
-[MOTION.md](https://github.com/firejune/rigc/blob/main/docs/MOTION.md); this file says when to open it and what it will
+[MOTION.md](https://github.com/rig-c/core/blob/main/docs/MOTION.md); this file says when to open it and what it will
 not do for you.
 
 ## Non-negotiables
@@ -31,16 +31,16 @@ thing that judges a movement is a person's eye through `rigc vote` — MOTION §
 
 ## Read, in this order
 
-1. [AUTHORING.md](https://github.com/firejune/rigc/blob/main/docs/AUTHORING.md) — the motion spec field by field (§4),
+1. [AUTHORING.md](https://github.com/rig-c/core/blob/main/docs/AUTHORING.md) — the motion spec field by field (§4),
    the failure map (§5–§6), reading reference frames (§8) and checking against
    them (§9), what the editor does when nobody tells it otherwise (§10), and
    reading a pose out of a picture (§11–§12).
-2. [MOTION.md](https://github.com/firejune/rigc/blob/main/docs/MOTION.md) — the normal form every motion request
+2. [MOTION.md](https://github.com/rig-c/core/blob/main/docs/MOTION.md) — the normal form every motion request
    reduces to (§0), timing, easing and the per-bone offset table (§3), and how to
    spread candidates so a ballot informs (§4–§5).
-3. Then [RIGGING.md](https://github.com/firejune/rigc/blob/main/docs/RIGGING.md) if the skeleton itself is what you have
-   to decide; [FACE.md](https://github.com/firejune/rigc/blob/main/docs/FACE.md) if the movement is a blink, a gaze or a
-   head turn; [INGEST.md](https://github.com/firejune/rigc/blob/main/docs/INGEST.md) if the rig arrived as a compiled
+3. Then [RIGGING.md](https://github.com/rig-c/core/blob/main/docs/RIGGING.md) if the skeleton itself is what you have
+   to decide; [FACE.md](https://github.com/rig-c/core/blob/main/docs/FACE.md) if the movement is a blink, a gaze or a
+   head turn; [INGEST.md](https://github.com/rig-c/core/blob/main/docs/INGEST.md) if the rig arrived as a compiled
    skeleton rather than loose parts.
 
 Every guide linked here is in the installed package at `node_modules/rig-c/docs/`,