spine-rigc 0.5.0 → 0.7.0

package/docs/SPEC_COVERAGE.md

@@ -5,18 +5,20 @@ what rigc emits today, and what the official example projects actually use — s
 ordered by what the benchmark ladder needs rather than by the spec's alphabet.
 
 Reproduce the measurements with `bun run fetch-examples && bun run bench:usage`
-([`bench/count_features.ts`](../bench/count_features.ts) → [`feature_matrix.csv`](feature_matrix.csv),
-[`feature_matrix.json`](feature_matrix.json)).
+([`bench/count_features.ts`](https://github.com/firejune/rigc/blob/main/bench/count_features.ts) → [`feature_matrix.csv`](https://github.com/firejune/rigc/blob/main/docs/feature_matrix.csv),
+[`feature_matrix.json`](https://github.com/firejune/rigc/blob/main/docs/feature_matrix.json)).
 
 > 🔼 **This note is dated and does not move; the measurements below are the corpus as it
 > stood on 2026-08-22 and stay as written. Its statements about what rigc *does* have gone
-> stale in five places, and the live ledger is [LADDER.md](LADDER.md):**
+> stale in five places, and the live ledger is [LADDER.md](https://github.com/firejune/rigc/blob/main/docs/LADDER.md):**
 >
 > - **B2 is fixed** — A16 accepts `4.3`, `4.3.N` and `4.3.N-<suffix>`, so all twelve example
 >   exports pass it.
-> - **The profile split exists** — `--profile spine | spine-html`, default `spine-html`, so the
->   nine renderer-policy assertions of §2.2 no longer apply to foreign data. That closes the
->   validator half of B3; the emitter half (no packer, no atlas importer) is untouched.
+> - **The profile split exists** — `--profile spine | spine-html`, **default `spine` since
+>   2026-08-29** (issue #221; it was `spine-html` when this survey was written), so the nine
+>   renderer-policy assertions of §2.2 no longer apply to foreign data, and no longer apply
+>   unasked. That closes the validator half of B3; the emitter half (no packer, no atlas
+>   importer) is untouched.
 > - **`walkTimelines` reaches all eleven groups** — §2.3 item 1 and §4.3 item 4 are done, so
 >   A05 is no longer blind to `ik` / `transform` / `path` / `slider` / `drawOrder` /
 >   `drawOrderFolder` / `events`.
@@ -25,7 +27,7 @@ Reproduce the measurements with `bun run fetch-examples && bun run bench:usage`
 >   attachments, the 4.3 typed constraint array and the invariants, all as data. So §2.4's
 >   "the bone tree is code, not data", §4.0's B1 row and executive-summary item 7 describe a
 >   state that no longer exists. Rung 3 transcribed into that format scores 1.000 on every
->   measure of every section — figures in [LADDER.md](LADDER.md).
+>   measure of every section — figures in [LADDER.md](https://github.com/firejune/rigc/blob/main/docs/LADDER.md).
 > - Since that change rigc also emits **ik and transform constraints**, the **single-axis bone
 >   timelines** (`translatex`/`translatey`/`scalex`/`scaley`/`shear`/`shearx`/`sheary`), bone
 >   `length`/`scale`/`shear`/`inherit`/`skin`/`color` setup fields, slot `dark`/`blend`, region
@@ -659,7 +661,7 @@ This is the split Part 4(c) needs. **Spine-validity** = the file is wrong for an
 | `A12_NO_DARK_COLOR` | **renderer-profile** | slot `dark`, `rgba2`/`rgb2` timelines — "parsed, then ignored" |
 | `A13_MESH_BUDGET` | **renderer-profile** | >4 mesh slots, >80 triangles per mesh |
 | `A14_NO_FULL_FRAME_MESH` | **renderer-profile** | a mesh spanning the whole stage |
-| `A19_OVERLAY_PNGS_HAVE_ALPHA` | **renderer-profile** | an overlay page with no alpha channel |
+| `A19_OVERLAY_PNGS_HAVE_ALPHA` | **renderer-profile** | an overlay page that can never be transparent — no alpha channel and no `tRNS` chunk |
 | `A20_MESH_WEIGHTS_COHERENT` | **renderer-profile / archetype** | **unweighted meshes**, weights not summing to 1, out-of-range bone indices |
 | `A21_MESH_RIM_PINNED` | archetype | ring/ribbon rim not pinned |
 | `A22_MESH_UVS_IN_UNIT_RANGE` | validity-ish | UVs outside 0..1 |
@@ -716,7 +718,7 @@ one.
 
 > 🔼 Both halves of that paragraph are now stale. The walker reaches all eleven groups and the
 > mutants that prove it exist; the suite runs on fixtures it **generates**, so it needs no
-> project's art. Current shape: [CLAUDE.md](../CLAUDE.md), *The selftest and its fixtures*.
+> project's art. Current shape: [CLAUDE.md](https://github.com/firejune/rigc/blob/main/CLAUDE.md), *The selftest and its fixtures*.
 
 ### 2.4 The input model — where the rig actually comes from
 
@@ -751,10 +753,10 @@ The manifest supplies *positions* for those bones (`anchors`), never their *name
 ## Part 3 — what the example ladder actually uses
 
 Corpus: the nine official example projects on branch `4.3`, fetched by
-[`scripts/fetch-examples.sh`](../scripts/fetch-examples.sh) into the gitignored `examples/<name>/`
+[`scripts/fetch-examples.sh`](https://github.com/firejune/rigc/blob/main/scripts/fetch-examples.sh) into the gitignored `examples/<name>/`
 (**12 skeleton JSON files** — some examples ship more than one — **10 `.atlas` files**, 8 `license.txt`).
-Counted by [`bench/count_features.ts`](../bench/count_features.ts); full output in
-[`feature_matrix.json`](feature_matrix.json) and [`feature_matrix.csv`](feature_matrix.csv)
+Counted by [`bench/count_features.ts`](https://github.com/firejune/rigc/blob/main/bench/count_features.ts); full output in
+[`feature_matrix.json`](https://github.com/firejune/rigc/blob/main/docs/feature_matrix.json) and [`feature_matrix.csv`](https://github.com/firejune/rigc/blob/main/docs/feature_matrix.csv)
 (12 rows × 135 columns), both regenerated by `bun run bench:usage`.
 
 > ⚠️ **Verification note.** The timeline columns below were re-derived independently from the raw JSON
@@ -913,7 +915,7 @@ Crossing Part 2 with Part 3. Categories, as requested:
 **(d) JSON-inexpressible** — things the format genuinely cannot hold.
 
 **Nothing below was implemented when this was written. This is a survey.** Three of its items
-have been built since — see the banner at the top of this document and [LADDER.md](LADDER.md) for
+have been built since — see the banner at the top of this document and [LADDER.md](https://github.com/firejune/rigc/blob/main/docs/LADDER.md) for
 which; the requirement text itself is left exactly as surveyed.
 
 ### 4.0 Three blockers that precede rung 1
@@ -977,7 +979,7 @@ Nothing structurally new (attachment timeline, blend, inherit all arrived by run
 #### Rung 6 — `6-arcs`
 
 > ✅ **Measured, 2026-08-23.** The rows below were written as predictions; the
-> transcription in [`bench/transcriptions/6-arcs/`](../bench/transcriptions/6-arcs/)
+> transcription in [`bench/transcriptions/6-arcs/`](https://github.com/firejune/rigc/tree/main/bench/transcriptions/6-arcs/)
 > settled them. It compiles green under `--profile spine` and `bench 6` reports
 > **1.000 on all 44 measures in all six sections**, so the rig spec expresses this
 > skeleton — weighted meshes, `edges` and the 4.3 transform constraints alike. A

package/package.json

@@ -1,10 +1,10 @@
 {
   "name": "spine-rigc",
-  "version": "0.5.0",
+  "version": "0.7.0",
   "description": "Rig compiler for Spine — declarative rig specs in, Spine 4.3 skeleton data out, verified by a spine-core round-trip. Built so AI agents can author rigs and check their own work; the output imports into the Spine editor.",
   "type": "module",
   "bin": {
-    "rigc": "./cli.ts"
+    "rigc": "./bin/rigc.cjs"
   },
   "engines": {
     "bun": ">=1.2.0"
@@ -31,12 +31,14 @@
     "cli"
   ],
   "files": [
+    "bin/rigc.cjs",
     "cli.ts",
     "src",
     "tools/plate.ts",
     "tools/font5x7.ts",
     "docs/AUTHORING.md",
     "docs/SPEC_COVERAGE.md",
+    "docs/PROMPTING.md",
     "NOTICE.md"
   ],
   "publishConfig": {