feat(pipeline): auto-generate embeddings during full indexing

When CBM_EMBEDDING_URL is configured, the full indexing pipeline now
automatically generates semantic embeddings after process and channel
detection. Uses force=false so existing embeddings are preserved —
only new/missing nodes get embedded.

This makes semantic search zero-friction: repos indexed while the
embedding server is running get embeddings automatically. No manual
generate_embeddings call needed.

Pipeline order: FTS5 backfill → processes → channels → embeddings

Author: Your Name

src/pipeline/pipeline.c

@@ -12,8 +12,8 @@
  */
 #include "pipeline/pipeline.h"
 #include "pipeline/pipeline_internal.h"
-// NOLINTNEXTLINE(misc-include-cleaner) — worker_pool.h included for interface contract
 #include "pipeline/worker_pool.h"
+#include "pipeline/embedding.h"
 #include "graph_buffer/graph_buffer.h"
 #include "store/store.h"
 #include "discover/discover.h"
@@ -856,6 +856,24 @@ int cbm_pipeline_run(cbm_pipeline_t *p) {
                     cbm_store_close(ch_store);
                 }
             }
+
+            /* ── Embedding generation: semantic vectors for hybrid search ── */
+            if (cbm_embedding_is_configured()) {
+                cbm_store_t *emb_store = cbm_store_open_path(db_path);
+                if (emb_store) {
+                    int nemb = cbm_embedding_generate_for_project(
+                        emb_store, p->project_name, false /* don't force — skip existing */);
+                    if (nemb > 0) {
+                        cbm_log_info("pass.done", "pass", "embeddings",
+                                     "generated", itoa_buf(nemb));
+                    } else if (nemb == 0) {
+                        cbm_log_info("pass.skip", "pass", "embeddings",
+                                     "reason", "all_exist");
+                    }
+                    /* nemb < 0 means error — already logged inside */
+                    cbm_store_close(emb_store);
+                }
+            }
         }
     }