Sync ako/mxcli: microflow rewrite carries, run --local resolution, version gating, catalog entry-point edges

.claude/commands/mxcli-dev/fix-issue.md

@@ -0,0 +1,88 @@
+---
+description: Diagnose and fix a reported bug, with the evidence bar the PR checklist expects
+---
+
+# /mxcli-dev:fix-issue — Fix a Reported Bug
+
+Work a bug report end to end: match it against what has already been seen, fix it,
+and prove the fix with evidence a reviewer can check.
+
+Read [`.claude/skills/fix-issue.md`](../../skills/fix-issue.md) first — it holds the
+findings-lookup mechanics and the two rules that are easiest to skip. This command
+is the order to do things in.
+
+`$ARGUMENTS` is the issue number or a description of the symptom.
+
+## Steps
+
+1. **Read the issue in full**, including comments. Quote the reported symptom
+   verbatim somewhere — it is what the test has to reproduce, and paraphrasing it
+   is how a fix ends up addressing a different bug.
+
+2. **Match the failure class**, then the instance:
+   - `docs-wiki/bug-patterns/` for the class (small, read it)
+   - `grep -il '<CE code or keyword>' .claude/skills/fix-issue/findings/*.jsonl`
+     for the instance
+
+   A pattern-page miss means "not yet digested", never "not seen before".
+
+3. **Establish the version facts before theorising**, if the report involves one.
+   The arbiter for a metamodel property is the Mendix Model SDK's own
+   `StructureVersionInfo` (`npm pack mendixmodelsdk`, then grep `src/gen/*.js` for
+   the type) — not release notes, and not a number already written down in this
+   repo. mendixlabs/mxcli#1121 was a floor copied from a proposal's illustrative
+   sample output that nobody had ever measured.
+
+4. **Write the failing test first**, at the layer the symptom lives in. Table of
+   layer → package is in the skill.
+
+5. **Prove the test detects the bug.** Revert the fix, or stub the guard, and
+   confirm it fails *with the reported symptom*. Put the control's output in the PR.
+   A test that has only ever run against fixed code has not been shown to detect
+   anything.
+
+6. **Run the real thing when the argument calls for it.** Required when the fix's
+   justification asserts what a Mendix tool accepts or rejects — a version guard, a
+   refusal, anything resting on "mxbuild would catch this". Also when the symptom is
+   a property of the running app (`verify-in-runtime.md`). Cheapest form:
+
+   ```bash
+   mxcli new <Name> --version <full version> --theme none --layout none --skip-init
+   mxcli exec <repro>.mdl -p <Name>.mpr
+   mxcli docker check -p <Name>.mpr          # and again with the fault forced back in
+   mxcli run --local -p <Name>.mpr           # when it has to render
+   ```
+
+   Build **two** copies — fixed, and with the fault forced back in — or the run
+   tells you nothing you did not already believe.
+
+7. **Add the regression case**: `mdl-examples/bug-tests/<issue>-<description>.mdl`,
+   and check it parses (`mxcli check`) and passes `make check-mdl`.
+
+8. **Append one finding** to `.claude/skills/fix-issue/findings/<area>.jsonl` and run
+   `make check-findings`. Write the insight — what would have made this cheaper to
+   find, and which plausible wrong turn to skip — not the changelog.
+
+9. `make build && make test && make lint`, then commit.
+
+## Before you say it cannot be verified
+
+That claim ends the investigation, so it needs the evidence a fix would. State what
+a positive result would look like; try the case you are sure works as a control; try
+a different *shape* of query rather than another value. Then, if it still holds, say
+it with the evidence attached rather than as a property of the environment.
+
+On #1121 the claim was "Mendix 10 is not downloadable here", from a dozen uniform
+404s with 11.x succeeding on the same host. It was wrong — Mendix 9 and 10 publish
+four-part names with a build number — and a uniform negative across a whole class
+was the tell that the query, not the class, was at fault.
+
+## Done when
+
+- [ ] Reported symptom reproduced by a test before the fix existed
+- [ ] Control run recorded (reverted fix → test fails with that symptom)
+- [ ] Full run done, with both variants, if the argument asserts what a tool accepts
+- [ ] Any "cannot verify" claim carries its evidence and a falsifying control
+- [ ] Bug-test MDL committed; `make check-mdl` passes
+- [ ] Finding appended; `make check-findings` passes
+- [ ] `make build && make test && make lint` pass

.claude/commands/mxcli-dev/proposal.md

@@ -25,7 +25,7 @@ Ask the user:
 
 If the user isn't sure about version or BSON, help them find out:
 - Version: check `reference/mendixmodellib/reflection-data/` or Mendix release notes
-- BSON: check if similar features exist in `sdk/mpr/parser*.go` or `sdk/mpr/writer*.go`
+- BSON: check if similar features exist in `mdl/backend/modelsdk/*_read.go` or `*_write.go`
 
 ### Phase 2: BSON Investigation (if applicable)
 
@@ -69,7 +69,7 @@ Before writing the proposal, search for existing work:
 ls docs/11-proposals/ | grep -i <feature>
 
 # Existing implementations
-grep -r "<feature>" mdl/executor/ sdk/mpr/ --include="*.go" -l
+grep -r "<feature>" mdl/executor/ mdl/backend/modelsdk/ --include="*.go" -l
 
 # Existing test coverage
 ls mdl-examples/doctype-tests/ | grep -i <feature>

.claude/commands/mxcli-dev/review.md

@@ -51,6 +51,8 @@ proactively. Add a row after every review that surfaces something new.
 | 21 | A registry that reads only one source (an `embed.FS`) has no extension point, so the only way to customise its output is to hand-edit the generated block — which a digest fence then refuses to touch on the next run. The feature ends up hostile to the exact case it was built for | API design | When a guard refuses a user's edit, ask whether the thing being edited should have been authorable. Introduce a source abstraction (`fs.FS` + root) and let a project-local directory shadow the embedded set — the walk functions usually already take a root, so the change is contained |
 | 22 | A test helper that expands a template cross-products every list against every reference (e.g. every `@each $weight` list against every `url()`), inventing artefacts no asset ever shipped and failing on correct input | Test coverage | Scan positionally: a reference must be expanded against the loop it actually sits under. Prove the helper both fires on a real break (delete one shipped file) and stays quiet on correct input — a helper only checked against green code has not been shown to detect anything |
 | 23 | A copy-to-scaffold path renames files but not the identifiers built from the name (`@mixin mxcli-<name>-<alt>`, `@import "mxcli-<name>"`), so two artefacts collide the moment both exist — and the symptom is a rule that silently compiles to nothing | Code correctness | Assert the structural contract on the *generated* artefact, not just the shipped ones: factor the built-in's contract test into a helper and run the scaffold through it. Verify once end to end against the real toolchain and record it |
+| 24 | A skill or command still instructs work into a package the repo has deleted (`sdk/mpr` test locations, a symptom table moved to `findings/*.jsonl` years prior) — the doc reads as authoritative and every instruction in it is a compile error or a no-op | Docs quality | When a package is deleted or a doc is restructured, grep `.claude/` for its name in the same PR. A deletion that leaves the guidance behind is worse than no guidance, because the reader trusts it |
+| 25 | A doc or skill shows a CLI invocation nobody ran — a command name that does not exist (`mxcli dump-bson` for `mxcli bson dump`), or a flag form the parser rejects (`--compare "A" "B"` where `--compare` is a StringSlice needing `"A,B"`). Worst when copied FROM the command's own `--help`, which had the same error, so the doc looks sourced | Docs quality | Run every command a doc shows, against a real project, before committing it. If it came from `--help`, run that form too — the example in the help text is not evidence that it works |
 
 ---
 

.claude/lint-rules/orphaned_elements.star

@@ -28,7 +28,7 @@ ENTRY_PAGE_PATTERNS = ["Home", "Login", "Index", "Dashboard"]
 # Reference kinds that mean "something causes this microflow to run". These are
 # catalog RefKind values (mdl/catalog/builder_references.go); a kind missing here
 # turns a live document into a false "not called from anywhere" finding.
-MICROFLOW_ENTRY_KINDS = ["call", "schedule", "datasource", "action", "calculate", "settings"]
+MICROFLOW_ENTRY_KINDS = ["call", "schedule", "publish", "datasource", "action", "calculate", "settings", "event"]
 
 # Reference kinds that mean "something opens this page".
 PAGE_ENTRY_KINDS = ["show_page", "home_page", "login_page", "menu_item", "action"]