chore: align repo scaffolding with the package standard

.github/workflows/ci.yml

@@ -1,38 +1,28 @@
-name: CI
+name: ci
 
 on:
+  pull_request:
   push:
     branches: [main]
-  pull_request:
 
 jobs:
   check:
     runs-on: ubuntu-latest
-    strategy:
-      matrix:
-        node-version: [24]
+    timeout-minutes: 15
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v5
       - uses: actions/setup-node@v4
         with:
-          node-version: ${{ matrix.node-version }}
+          node-version: 24
       - uses: oven-sh/setup-bun@v2
-        with:
-          bun-version: latest
       - run: bun install --frozen-lockfile
-      - run: bunx tsc --noEmit
-      - run: bun run lint
-      - run: bun run test
+      - run: bun run check
       - run: bun run test:e2e
-      - run: bun run format:check
-      - run: bun run build
-      - name: Import packed dist with Node
+      - name: node consumer smoke
         run: |
           set -euo pipefail
-          tarball=$(npm pack --silent | tail -n 1)
-          tmpdir=$(mktemp -d)
-          trap 'rm -rf "$tmpdir" "$tarball"' EXIT
-          cd "$tmpdir"
-          npm init -y >/dev/null
-          npm install "$OLDPWD/$tarball"
-          node --input-type=module -e "import { embedTexts } from '@corbits/embedding'; if (typeof embedTexts !== 'function') process.exit(1)"
+          TARBALL="$PWD/$(npm pack --silent)"
+          mkdir -p "$RUNNER_TEMP/c" && cd "$RUNNER_TEMP/c"
+          npm init -y >/dev/null && npm pkg set type=module >/dev/null
+          npm install "$TARBALL"
+          node -e 'import("@corbits/embedding").then((m) => { for (const n of ["embedTexts", "probeEmbedDims", "EmbeddingRequestError"]) if (typeof m[n] !== "function") throw new Error("missing export: " + n); })'

.github/workflows/cla.yml

@@ -0,0 +1,53 @@
+name: CLA Assistant
+
+on:
+  issue_comment:
+    types: [created]
+  pull_request_target:
+    types: [opened, synchronize]
+
+permissions:
+  actions: write
+  contents: write
+  pull-requests: write
+  statuses: write
+
+concurrency:
+  group: cla-${{ github.event.pull_request.number || github.event.issue.number || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  cla:
+    runs-on: ubuntu-latest
+    timeout-minutes: 10
+    steps:
+      - name: Maintainer fast path
+        if: >-
+          github.event_name == 'pull_request_target' &&
+          (contains(fromJSON('["TheGreatAxios","brianjfox"]'), github.event.pull_request.user.login) ||
+           endsWith(github.event.pull_request.user.login, '[bot]'))
+        run: echo "Maintainer or bot pull request; CLA not required."
+      - name: CLA Assistant
+        if: >-
+          ((github.event.comment.body == 'recreate-signatures' ||
+            github.event.comment.body == 'I have read the CLA Document and I hereby sign the CLA') ||
+           github.event_name == 'pull_request_target') &&
+          !(github.event_name == 'pull_request_target' &&
+            (contains(fromJSON('["TheGreatAxios","brianjfox"]'), github.event.pull_request.user.login) ||
+             endsWith(github.event.pull_request.user.login, '[bot]')))
+        # corbitsdev/cla-assistant-action v2.6.1-node24: upstream v2.6.1 on node24.
+        uses: corbitsdev/cla-assistant-action@ef6d3e51db8232fe93090810f13bde30497c1d74
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          path-to-document: "https://github.com/${{ github.repository }}/blob/main/CLA.md"
+          path-to-signatures: "signatures/version1/cla.json"
+          branch: "cla-signatures"
+          allowlist: TheGreatAxios,brianjfox,*[bot]
+          custom-notsigned-prcomment: >-
+            Thank you for your contribution. Before it can be merged, please read our
+            [Contributor License Agreement](https://github.com/${{ github.repository }}/blob/main/CLA.md)
+            and sign it by posting a new comment on this pull request containing
+            exactly the line below (nothing else):
+          custom-pr-sign-comment: "I have read the CLA Document and I hereby sign the CLA"
+          custom-allsigned-prcomment: "All contributors have signed the CLA."

.gitignore

@@ -1,3 +1,8 @@
-node_modules
-dist
+node_modules/
+dist/
 *.tsbuildinfo
+*.tgz
+coverage/
+.env
+.env.*
+.DS_Store

AGENTS.md

@@ -0,0 +1,39 @@
+# AGENTS.md
+
+## Purpose
+
+`@corbits/embedding` turns texts into vectors over any OpenAI-compatible
+`/v1/embeddings` endpoint, with `@intx/inference` retries and error
+classification. It owns batching, reply integrity and the one-shot JSON
+transport. It does not store or search vectors; that is `@corbits/memory`.
+
+## Layout
+
+- `src/embed.ts`: `EmbedConfigSchema`, `embedTexts`, `probeEmbedDims`, batch
+  splitting and reply parsing (float and base64).
+- `src/request.ts`: `runJSONRequest` (private transport), `EmbeddingRequestError`,
+  `Retry-After` extraction.
+- `src/index.ts`: the only module consumers import from.
+- `e2e/`: harness-driven retry, wire and base64 suites, plus a live round trip
+  gated on `EMBEDDING_E2E_BASE_URL`.
+
+## Rules
+
+- Place vectors by the echoed `index`, never by array position. Reject short,
+  duplicate or out-of-range replies instead of mispairing.
+- Batches run sequentially; retry policy is per request.
+- Throw `EmbeddingRequestError` carrying the classified `InferenceError` as
+  `reason`; never throw the `InferenceError` itself.
+- Read a 200 as text, then `JSON.parse`, so a non-JSON body classifies as a
+  protocol mismatch.
+- Omit `dimensions` and `encoding_format` from the wire unless configured.
+- Backoff time comes from `deps.scheduler`, not wall-clock globals.
+- `exactOptionalPropertyTypes` is on: omit optional keys, never assign
+  `undefined` to them.
+
+## Local development
+
+```sh
+bun install
+bun run check
+```

CLA.md

@@ -0,0 +1,64 @@
+# Contributor License Agreement (CLA)
+
+Version 1.0
+
+This Contributor License Agreement ("Agreement") is entered into between the contributor ("Contributor") and ABK Labs, Inc. ("Maintainer").
+
+By submitting any Contribution to a project maintained by Maintainer, the Contributor agrees to the following terms.
+
+## 1. Definitions
+
+"Contribution" means any source code, object code, documentation, test cases, designs, specifications, bug fixes, enhancements, comments, pull requests, commits, issues containing code, or other materials intentionally submitted to a project maintained by Maintainer.
+
+## 2. Copyright Ownership
+
+The Contributor retains ownership of the Contributor's copyrights in the Contribution.
+
+No transfer of copyright ownership is required by this Agreement.
+
+## 3. License Grant to Maintainer
+
+The Contributor grants Maintainer a perpetual, worldwide, non-exclusive, irrevocable, royalty-free license to:
+
+- use, reproduce, modify, distribute, display, perform, and sublicense the Contribution;
+- combine the Contribution with other software and works;
+- distribute the Contribution under the project's current license;
+- distribute the Contribution under future versions of the project's license;
+- distribute the Contribution under alternative open-source, source-available, commercial, proprietary, or dual-license terms.
+
+## 4. License Grant to the Public
+
+The Contributor agrees that the Contribution may be distributed as part of the applicable project under the project's then-current license terms, including LGPL-2.1 or any successor license adopted by Maintainer.
+
+## 5. Contributor Representations
+
+The Contributor represents and warrants that:
+
+1. The Contributor created the Contribution or otherwise has sufficient rights to submit it.
+2. The Contributor has the legal authority to grant the rights described in this Agreement.
+3. To the best of the Contributor's knowledge, the Contribution does not knowingly infringe the intellectual property rights of any third party.
+4. Any third-party material included in the Contribution has been properly disclosed and is compatible with the rights granted under this Agreement.
+
+## 6. Corporate Contributions
+
+If a Contribution is submitted on behalf of an employer or other legal entity, the person accepting this Agreement represents that they are authorized to bind that entity to this Agreement.
+
+## 7. No Obligation
+
+Maintainer is not obligated to use, distribute, maintain, support, or accept any Contribution.
+
+## 8. Disclaimer
+
+Except as expressly stated in this Agreement, the Contribution is provided "AS IS" without warranties or conditions of any kind.
+
+## 9. Electronic Acceptance
+
+The Contributor agrees that electronic acceptance of this Agreement, including acceptance through a website, source-control platform, click-through process, pull-request workflow, or similar mechanism, shall have the same force and effect as a handwritten signature.
+
+## 10. Governing Law
+
+This Agreement shall be governed by the laws of the State of California, excluding its conflict-of-law provisions.
+
+---
+
+By submitting a Contribution, the Contributor acknowledges that they have read and agree to this Agreement.

CONTRIBUTING.md

@@ -2,33 +2,40 @@
 
 ## Development
 
-```bash
+```sh
 bun install
-bun run typecheck
-bun run lint
-bun run test
-bun run test:e2e
+bun run check
 ```
 
-`bun run test` runs the colocated units in `src/`. `bun run test:e2e` runs
-`e2e/`: retry, wire and base64 cases on the `@intx/inference-testing` harness
-and a local HTTP stub, plus a `/v1/embeddings` round trip that skips unless
-`EMBEDDING_E2E_BASE_URL` is set; point it at a remote server rather than
-loading models locally:
+`bun run check` runs typecheck, lint, format check and unit tests. `bun run format` rewrites the tree.
 
-```bash
-EMBEDDING_E2E_BASE_URL=http://100.113.184.123:11434/v1 bun run test:e2e
-```
-
-`EMBEDDING_E2E_MODEL` defaults to `nomic-embed-text`.
+Contributors sign the [CLA](CLA.md) on their first PR; the CLA bot explains how.
 
-## Versioning
+`bun run test:e2e` runs `e2e/`: retry, wire and base64 cases on the
+`@intx/inference-testing` harness and a local HTTP stub, plus a
+`/v1/embeddings` round trip that skips unless `EMBEDDING_E2E_BASE_URL` is set
+(`EMBEDDING_E2E_MODEL` defaults to `nomic-embed-text`):
 
-Semver. Releases run `bun run build && npm publish` with green CI.
+```sh
+EMBEDDING_E2E_BASE_URL=http://<host>:11434/v1 bun run test:e2e
+```
 
 ## Commit messages
 
 Commit subjects and PR titles follow [Conventional Commits](https://www.conventionalcommits.org): `feat`, `fix`, `refactor`, `test`, `docs`, `build`, `ci`, `perf`, and `chore(release): x.y.z` for releases.
 Add `!` only for public API breaks: removed or renamed exports, changed signatures, newly required params. Peer and dependency range changes are `build(deps):` with no `!`.
 Keep subjects imperative, lowercase after the colon, 72 characters or less, and free of ticket IDs.
 Every PR links its issue with a `Closes <issue id>` line in the PR body.
+
+## Releasing
+
+Releases are manual. On a clean, up-to-date `main`:
+
+```sh
+npm version <patch|minor> -m "chore(release): %s"
+git push --follow-tags
+gh release create "v$(node -p 'require("./package.json").version')" --generate-notes
+npm publish
+```
+
+Bump minor only for breaking API changes; everything else is a patch. `prepack` builds `dist/` from the tagged commit.