fix: prevent fuzzy dedup from collapsing distinct same-named symbols on --update (#1178)

Three-part fix:

dedup.py: Pass 1 exact-merge now skips nodes with an empty source_file.
Previously all no-source_file nodes with the same label landed in one
bucket and were merged, destroying distinct symbols (third-party deps,
standalone functions) that happened to share a short name.

update.md (skillgen + all 13 host variants): the --update merge now
passes both deleted AND changed files to prune_sources, mirroring what
watch._rebuild_code already does correctly. Old nodes for re-extracted
files are pruned before fresh AST is inserted — no fuzzy reconciliation
needed, no cross-file collapse possible.

export.py: anti-shrink guard message now names fuzzy dedup as a
possible cause (not only "missing chunk files"), and advises a full
rebuild as the safe recovery path.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

graphify/dedup.py

@@ -186,7 +186,11 @@ def deduplicate_entities(
         for node in group:
             sf = node.get("source_file") or ""
             by_file[sf].append(node)
-        for file_group in by_file.values():
+        for sf, file_group in by_file.items():
+            if not sf:
+                # No source_file — cannot prove same symbol; skip to avoid
+                # collapsing distinct nodes that happen to share a label (#1178).
+                continue
             if len(file_group) > 1:
                 winner = _pick_winner(file_group)
                 for node in file_group:

graphify/export.py

@@ -493,9 +493,12 @@ def to_json(G: nx.Graph, communities: dict[int, list[str]], output_path: str, *,
                 import sys as _sys
                 print(
                     f"[graphify] WARNING: new graph has {new_n} nodes but existing "
-                    f"graph.json has {existing_n}. Refusing to overwrite — you may be "
-                    f"missing chunk files from a previous session. "
-                    f"Pass force=True to override.",
+                    f"graph.json has {existing_n} (net -{existing_n - new_n}). "
+                    f"Refusing to overwrite. Possible causes: missing chunk files from "
+                    f"a previous session, or fuzzy dedup collapsed same-named symbols "
+                    f"across files during an --update on an already-current graph. "
+                    f"Run a full rebuild (/graphify .) to be safe, or pass force=True "
+                    f"only if you have verified the reduction is legitimate.",
                     file=_sys.stderr,
                 )
                 return False

graphify/skills/amp/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/claude/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/claw/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/codex/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/copilot/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/droid/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/kilo/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/kiro/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/opencode/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/pi/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/trae/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/vscode/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

graphify/skills/windows/references/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__amp__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__claude__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__claw__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__codex__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__copilot__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__droid__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__kilo__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__kiro__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__opencode__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__pi__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__trae__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__vscode__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/expected/graphify__skills__windows__references__update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')
 

tools/skillgen/fragments/references/shared/update.md

@@ -93,13 +93,18 @@ from graphify.detect import save_manifest
 new_extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
 incremental = json.loads(Path('graphify-out/.graphify_incremental.json').read_text(encoding=\"utf-8\"))
 deleted = list(incremental.get('deleted_files', []))
+# Also prune old nodes for re-extracted (changed) files before inserting fresh AST.
+# Without this, build_merge's dedup pass tries to reconcile old and new versions of
+# the same file's nodes and can collapse same-named symbols across files (#1178).
+changed = [f for files in incremental.get('new_files', {}).values() for f in files]
+prune = list(dict.fromkeys(deleted + changed)) or None
 
 # Use build_merge() — reads graph.json directly without NetworkX round-trip
 # so edge direction (calls, implements, imports) is always preserved (#801).
 G = build_merge(
     [new_extraction],
     graph_path='graphify-out/graph.json',
-    prune_sources=deleted or None,
+    prune_sources=prune,
 )
 print(f'[graphify update] Merged: {G.number_of_nodes()} nodes, {G.number_of_edges()} edges')