Release v0.23.0: package the English core workflow

.github/workflows/release.yml

@@ -195,18 +195,24 @@ jobs:
         shell: bash
         run: |
           set -euo pipefail
+          notes="docs/releases/$TAG.md"
+          if [[ -f "$notes" ]]; then
+            note_args=(--notes-file "$notes")
+          else
+            note_args=(--generate-notes)
+          fi
           if [[ "$EXISTING_RELEASE" == "true" ]]; then
             gh release upload "$TAG" dist/* --clobber
           elif [[ "$EXISTING_TAG" == "true" ]]; then
             gh release create "$TAG" dist/* \
               --verify-tag \
               --title "AgentConfigScore $TAG" \
-              --generate-notes
+              "${note_args[@]}"
           else
             gh release create "$TAG" dist/* \
               --target "$GITHUB_SHA" \
               --title "AgentConfigScore $TAG" \
-              --generate-notes
+              "${note_args[@]}"
           fi
 
       - name: Update rolling v0 branch

CHANGELOG.md

@@ -2,6 +2,15 @@
 
 ## Unreleased
 
+No changes yet.
+
+## v0.23.0
+
+The English core workflow is packaged as a local deterministic linter and CI
+regression gate. This remains a Beta release, not a semantic quality or security
+certification. See the [release notes](docs/releases/v0.23.0.md) for installation,
+compatibility changes, and known limits.
+
 ### Added
 
 - An offline installed-package acceptance journey covering both console aliases and the module entrypoint, setup, reports, pass/fail regression behavior, baseline-policy bypass protection and operational failures. CI exercises the wheel on Linux, Windows and macOS; manual publication now waits for the full source suite and all installed acceptance jobs.

CITATION.cff

@@ -2,8 +2,8 @@ cff-version: 1.2.0
 message: "If you use AgentConfigScore in research, please cite this software."
 title: "AgentConfigScore: deterministic lint and regression checks for AI coding-agent instructions"
 type: software
-version: 0.22.0
-date-released: 2026-09-07
+version: 0.23.0
+date-released: 2026-10-07
 authors:
   - family-names: Lin
     given-names: Zhibo

README.md

@@ -609,7 +609,7 @@ The Action emits outputs before returning its final regression status. Use `if:
 - [x] installed-wheel end-to-end acceptance on Linux, Windows and macOS in CI
 - [x] source and installed-package gates before publication
 - [x] user guide, exit-code reference and troubleshooting
-- [ ] next package cut after version/changelog review
+- [x] versioned English core package with release notes and compatibility guidance
 - [ ] optional Simplified and Traditional Chinese documentation and README language switch
 - [ ] optional semantic contradiction plugin
 

docs/releases/v0.23.0.md

@@ -0,0 +1,74 @@
+# AgentConfigScore v0.23.0
+
+This release packages the English core workflow: a dependency-free, local
+instruction linter and baseline-aware CI regression gate. It remains **Beta**.
+No accuracy comparison with AI review or independent user validation is claimed.
+
+## Install or upgrade
+
+```bash
+python -m pip install --upgrade "agent-config-score==0.23.0"
+agent-config-score --version
+```
+
+Expected version: `AgentConfigScore 0.23.0`. Python 3.10+ is required; Git is
+needed for `diff`. Both console aliases (`agent-config-score`, `acs`) and the
+PATH-independent `python -m agent_config_score` entrypoint are available.
+
+## What is included
+
+- Shared Markdown context for more conservative handling of explicit prohibited
+  command blocks/lists, matching code fences and command-option parsing.
+- Local Markdown-link checks relative to the instruction file.
+- Narrow English rules for explicitly false success reports and bulk environment
+  transfers. These are evidence-based lint signals, not general intent detection.
+- Baseline-owned regression policy and expiring suppressions, local `diff` and
+  `compare`, initialization, integration doctor, rule inspection and reports.
+- An optional copy-ready PR-comment workflow; posting is opt-in, not an automatic
+  action of the scanner.
+- Windows Unicode Git-path fixes and concise diagnostics for invalid UTF-8
+  integration files or unwritable reports.
+- A [user guide](https://github.com/LE0-Lin/AgentConfigScore/blob/v0.23.0/docs/user-guide.md)
+  covering setup, exit codes and troubleshooting, plus a
+  [documentation index](https://github.com/LE0-Lin/AgentConfigScore/blob/v0.23.0/docs/README.md).
+- Installed-wheel acceptance on Linux, Windows and macOS, required before
+  publication alongside the full source regression suite.
+
+## Compatibility and adoption
+
+Existing command names, configuration version 1 and JSON field names remain
+unchanged. Rule additions and interpretation fixes can change scores and findings
+on the same input compared with v0.22.0. Review new signals when upgrading; a
+score change is not a new real-world accuracy measurement.
+
+The new stable rule IDs are `false-success-report` and `broad-environment-upload`.
+They are included in rule inspection, suppression schema and SARIF metadata.
+Operational filesystem/encoding failures return exit `2`; doctor error checks
+return `1`. These failures are not clean scan results.
+
+For reproducible CI, pin the Action to `LE0-Lin/AgentConfigScore@v0.23.0` (or its
+verified commit). `@v0` remains a rolling pre-1.0 reference and can change on a
+later release. The tool applies the baseline policy to both trees so a PR cannot
+use its own weaker policy or new suppression to approve itself.
+
+Existing generated workflows do not need to be reinitialized merely to upgrade
+the installed package. Review whether they use a fixed Action version or rolling
+`v0`, since those determine the CI tool separately from your local installation.
+
+## Evidence and known limits
+
+Maintained deterministic contracts and synthetic installed-package journeys
+verify expected engineering behavior; they do not establish real-world precision,
+recall or usefulness. The unresolved AI pilot and failed source acquisitions
+remain explicitly unscored; research preparation is separate from the runtime.
+
+An A 100 means no active deterministic rule matched, **not** semantic perfection
+or a security approval. Arbitrary paraphrases, contradictions and other-language
+instruction semantics still need review. Read the
+[limitations](https://github.com/LE0-Lin/AgentConfigScore/blob/v0.23.0/docs/limitations.md)
+and [full changelog](https://github.com/LE0-Lin/AgentConfigScore/blob/v0.23.0/CHANGELOG.md).
+
+Simplified and Traditional Chinese documentation/language links are the next
+optional milestone; they are **not included in v0.23.0**. Translating documentation
+will not by itself add Chinese semantic detection or change machine-readable
+interfaces.

docs/user-guide.md

@@ -16,10 +16,10 @@ agent-config-score --version
 agent-config-score --help
 ```
 
-`acs` is a shorter alias. Development source also supports
-`python -m agent_config_score` with the same product commands; this module
-entrypoint is not present in the already published v0.22.0. It will be included
-in the next package release. Check [releases](https://github.com/LE0-Lin/AgentConfigScore/releases)
+`acs` is a shorter alias. Version 0.23.0 and later also support
+`python -m agent_config_score` with the same product commands, without relying
+on the console script's PATH setup. Older versions do not provide this module
+entrypoint. Check [releases](https://github.com/LE0-Lin/AgentConfigScore/releases)
 before assuming a change on `main` is available on PyPI or the rolling `v0` ref.
 
 ## 2. Initialize in your repository
@@ -140,7 +140,7 @@ Nothing is uploaded. Review and sanitize the draft yourself before sharing.
 
 | Symptom | What to check |
 |---|---|
-| Command is not found | Install into the active Python environment; confirm its Scripts/bin directory is on PATH. The next package also provides `python -m agent_config_score`. |
+| Command is not found | Install into the active Python environment; confirm its Scripts/bin directory is on PATH. In v0.23.0+, use `python -m agent_config_score` as a PATH-independent alternative. |
 | Configuration / workflow encoding error | Save the file as UTF-8; do not copy a UTF-16 file into the integration. |
 | Cannot write HTML, SARIF or Markdown | Choose a writable file path, not an existing directory; inspect parent-directory permissions. |
 | `diff` cannot detect a baseline | Check `doctor`, then pass an available default-branch ref explicitly. |

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "agent-config-score"
-version = "0.22.0"
+version = "0.23.0"
 description = "Deterministic lint and regression gate for AGENTS.md, Claude, Cursor, Copilot, and Gemini instructions."
 readme = "README.md"
 requires-python = ">=3.10"

src/agent_config_score/__init__.py

@@ -1 +1 @@
-__version__ = "0.22.0"
+__version__ = "0.23.0"

tests/test_release_workflow.py

@@ -68,6 +68,24 @@ def test_privileged_third_party_publish_action_is_commit_pinned(self):
             workflow,
         )
 
+    def test_current_package_has_versioned_release_notes_and_changelog(self):
+        source = (ROOT / "src" / "agent_config_score" / "__init__.py").read_text(encoding="utf-8")
+        version = re.search(r'^__version__ = "([^"]+)"$', source, re.MULTILINE).group(1)
+        notes = ROOT / "docs" / "releases" / f"v{version}.md"
+        self.assertTrue(notes.is_file())
+        content = notes.read_text(encoding="utf-8")
+        self.assertIn(f"agent-config-score=={version}", content)
+        self.assertIn(f"AgentConfigScore {version}", content)
+        self.assertIn("Beta", content)
+        self.assertIn(f"## v{version}\n", (ROOT / "CHANGELOG.md").read_text(encoding="utf-8"))
+
+    def test_release_prefers_reviewed_notes_with_a_generated_fallback(self):
+        workflow = WORKFLOW.read_text(encoding="utf-8")
+        self.assertIn('notes="docs/releases/$TAG.md"', workflow)
+        self.assertIn('note_args=(--notes-file "$notes")', workflow)
+        self.assertIn("note_args=(--generate-notes)", workflow)
+        self.assertEqual(workflow.count('"${note_args[@]}"'), 2)
+
 
 if __name__ == "__main__":
     unittest.main()