diff --git a/internal/cbm/extract_imports.c b/internal/cbm/extract_imports.c
index 961244180..4e44d33ac 100644
--- a/internal/cbm/extract_imports.c
+++ b/internal/cbm/extract_imports.c
@@ -1,6 +1,7 @@
#include "cbm.h"
#include "arena.h" // CBMArena, cbm_arena_strdup/strndup/sprintf
#include "helpers.h"
+#include "lang_specs.h" // CBMLangSpec, CBMEmbeddedLangSpec, cbm_lang_spec, cbm_ts_language
#include "tree_sitter/api.h" // TSNode, ts_node_*
#include "foundation/constants.h"
#include "ts_node_stack.h"
@@ -846,6 +847,99 @@ static void parse_wolfram_imports(CBMExtractCtx *ctx) {
walk_wolfram_imports(ctx, ctx->root);
}
+// --- Embedded-language imports ---
+// Generic walker for host grammars (Svelte, Vue, HTML, Astro, ...) whose AST
+// keeps embedded sub-language source as raw_text (or similar) without parsing
+// it. The host's CBMLangSpec.embedded_imports declares which content nodes
+// hold which sub-language; we re-parse each match with the embedded grammar
+// and run the standard ES import walker over the inner AST.
+//
+// No grammar symbols are referenced here — the embedded TSLanguage is
+// resolved through cbm_ts_language(spec->embedded_language), the same hook
+// that the main parser uses. Adding another host language is a one-line
+// declaration in lang_specs.c.
+
+static void embedded_collect_content_nodes(TSNode root, const CBMEmbeddedLangSpec *spec,
+ TSNode *out, int *out_count, int max_out) {
+ /* Iterative DFS so deeply-nested script blocks are still found. Cap the
+ * stack to a sane bound (host grammars do not have million-deep markup
+ * trees) — no need to introduce TSNodeStack here. */
+ enum { EMBED_STACK_CAP = 1024 };
+ TSNode stack[EMBED_STACK_CAP];
+ int top = 0;
+ stack[top++] = root;
+ while (top > 0 && *out_count < max_out) {
+ TSNode node = stack[--top];
+ const char *kind = ts_node_type(node);
+ if (strcmp(kind, spec->script_node_type) == 0) {
+ uint32_t cc = ts_node_child_count(node);
+ for (uint32_t k = 0; k < cc; k++) {
+ TSNode c = ts_node_child(node, k);
+ if (strcmp(ts_node_type(c), spec->content_node_type) == 0) {
+ out[(*out_count)++] = c;
+ if (*out_count >= max_out) {
+ return;
+ }
+ break; /* one content node per script element */
+ }
+ }
+ /* Do not descend into \n"
+ "Hello {value}
\n",
+ CBM_LANG_SVELTE, "t", "Comp.svelte");
+ ASSERT_NOT_NULL(r);
+ ASSERT_FALSE(r->has_error);
+ ASSERT_GTE(r->imports.count, 3);
+ ASSERT(has_import(r, "Foo.svelte"));
+ ASSERT(has_import(r, "lib/utils"));
+ ASSERT(has_import(r, "helpers"));
+ cbm_free_result(r);
+ PASS();
+}
+
+TEST(svelte_imports_no_script) {
+ /* .svelte with no \n",
+ CBM_LANG_VUE, "t", "App.vue");
+ ASSERT_NOT_NULL(r);
+ ASSERT_FALSE(r->has_error);
+ ASSERT_GTE(r->imports.count, 2);
+ ASSERT(has_import(r, "MyComp.vue"));
+ ASSERT(has_import(r, "vue"));
+ cbm_free_result(r);
+ PASS();
+}
+
+TEST(html_imports_basic) {
+ /* Plain HTML with inline ES module imports — same generic walker. */
+ CBMFileResult *r = extract(
+ "\n"
+ "\n"
+ "\n",
+ CBM_LANG_HTML, "t", "index.html");
+ ASSERT_NOT_NULL(r);
+ ASSERT_FALSE(r->has_error);
+ ASSERT_GTE(r->imports.count, 2);
+ ASSERT(has_import(r, "app.js"));
+ ASSERT(has_import(r, "utils.js"));
+ cbm_free_result(r);
+ PASS();
+}
+
/* ═══════════════════════════════════════════════════════════════════
* config_extraction_test.go ports (25 tests)
* ═══════════════════════════════════════════════════════════════════ */
@@ -2297,6 +2378,10 @@ SUITE(extraction) {
RUN_TEST(ruby_imports);
RUN_TEST(lua_imports);
RUN_TEST(import_stress_go);
+ RUN_TEST(svelte_imports_basic);
+ RUN_TEST(svelte_imports_no_script);
+ RUN_TEST(vue_imports_basic);
+ RUN_TEST(html_imports_basic);
/* config_extraction_test.go ports */
RUN_TEST(toml_basic_table_and_pair);