diff --git a/graphify/skill-amp.md b/graphify/skill-amp.md
index de0dbe265..caf9ee798 100644
--- a/graphify/skill-amp.md
+++ b/graphify/skill-amp.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -308,7 +308,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -442,6 +442,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-claw.md b/graphify/skill-claw.md
index 6c7060a55..29c5f66a5 100644
--- a/graphify/skill-claw.md
+++ b/graphify/skill-claw.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-codex.md b/graphify/skill-codex.md
index f6080f3f3..c0232d006 100644
--- a/graphify/skill-codex.md
+++ b/graphify/skill-codex.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -308,7 +308,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -442,6 +442,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-copilot.md b/graphify/skill-copilot.md
index 6c7060a55..29c5f66a5 100644
--- a/graphify/skill-copilot.md
+++ b/graphify/skill-copilot.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-droid.md b/graphify/skill-droid.md
index 47323b613..8c343c5ef 100644
--- a/graphify/skill-droid.md
+++ b/graphify/skill-droid.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -308,7 +308,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -442,6 +442,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-kilo.md b/graphify/skill-kilo.md
index 2255c9166..089ac0df9 100644
--- a/graphify/skill-kilo.md
+++ b/graphify/skill-kilo.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-kiro.md b/graphify/skill-kiro.md
index 6c7060a55..29c5f66a5 100644
--- a/graphify/skill-kiro.md
+++ b/graphify/skill-kiro.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-opencode.md b/graphify/skill-opencode.md
index 893149ee5..776136b52 100644
--- a/graphify/skill-opencode.md
+++ b/graphify/skill-opencode.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -303,7 +303,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -437,6 +437,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-pi.md b/graphify/skill-pi.md
index 6c7060a55..29c5f66a5 100644
--- a/graphify/skill-pi.md
+++ b/graphify/skill-pi.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-trae.md b/graphify/skill-trae.md
index 01cccb1f8..ba9efcd9b 100644
--- a/graphify/skill-trae.md
+++ b/graphify/skill-trae.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -309,7 +309,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -443,6 +443,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-vscode.md b/graphify/skill-vscode.md
index 3016c581a..bb4d6a1b5 100644
--- a/graphify/skill-vscode.md
+++ b/graphify/skill-vscode.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -307,7 +307,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -441,6 +441,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill-windows.md b/graphify/skill-windows.md
index 2852de234..2642d9f5a 100644
--- a/graphify/skill-windows.md
+++ b/graphify/skill-windows.md
@@ -201,7 +201,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -246,7 +246,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -333,7 +333,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -467,6 +467,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/graphify/skill.md b/graphify/skill.md
index 6c7060a55..29c5f66a5 100644
--- a/graphify/skill.md
+++ b/graphify/skill.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-amp.md b/tools/skillgen/expected/graphify__skill-amp.md
index de0dbe265..caf9ee798 100644
--- a/tools/skillgen/expected/graphify__skill-amp.md
+++ b/tools/skillgen/expected/graphify__skill-amp.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -308,7 +308,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -442,6 +442,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-claw.md b/tools/skillgen/expected/graphify__skill-claw.md
index 6c7060a55..29c5f66a5 100644
--- a/tools/skillgen/expected/graphify__skill-claw.md
+++ b/tools/skillgen/expected/graphify__skill-claw.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-codex.md b/tools/skillgen/expected/graphify__skill-codex.md
index f6080f3f3..c0232d006 100644
--- a/tools/skillgen/expected/graphify__skill-codex.md
+++ b/tools/skillgen/expected/graphify__skill-codex.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -308,7 +308,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -442,6 +442,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-copilot.md b/tools/skillgen/expected/graphify__skill-copilot.md
index 6c7060a55..29c5f66a5 100644
--- a/tools/skillgen/expected/graphify__skill-copilot.md
+++ b/tools/skillgen/expected/graphify__skill-copilot.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-droid.md b/tools/skillgen/expected/graphify__skill-droid.md
index 47323b613..8c343c5ef 100644
--- a/tools/skillgen/expected/graphify__skill-droid.md
+++ b/tools/skillgen/expected/graphify__skill-droid.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -308,7 +308,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -442,6 +442,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-kilo.md b/tools/skillgen/expected/graphify__skill-kilo.md
index 2255c9166..089ac0df9 100644
--- a/tools/skillgen/expected/graphify__skill-kilo.md
+++ b/tools/skillgen/expected/graphify__skill-kilo.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-kiro.md b/tools/skillgen/expected/graphify__skill-kiro.md
index 6c7060a55..29c5f66a5 100644
--- a/tools/skillgen/expected/graphify__skill-kiro.md
+++ b/tools/skillgen/expected/graphify__skill-kiro.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-opencode.md b/tools/skillgen/expected/graphify__skill-opencode.md
index 893149ee5..776136b52 100644
--- a/tools/skillgen/expected/graphify__skill-opencode.md
+++ b/tools/skillgen/expected/graphify__skill-opencode.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -303,7 +303,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -437,6 +437,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-pi.md b/tools/skillgen/expected/graphify__skill-pi.md
index 6c7060a55..29c5f66a5 100644
--- a/tools/skillgen/expected/graphify__skill-pi.md
+++ b/tools/skillgen/expected/graphify__skill-pi.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-trae.md b/tools/skillgen/expected/graphify__skill-trae.md
index 01cccb1f8..ba9efcd9b 100644
--- a/tools/skillgen/expected/graphify__skill-trae.md
+++ b/tools/skillgen/expected/graphify__skill-trae.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -309,7 +309,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -443,6 +443,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-vscode.md b/tools/skillgen/expected/graphify__skill-vscode.md
index 3016c581a..bb4d6a1b5 100644
--- a/tools/skillgen/expected/graphify__skill-vscode.md
+++ b/tools/skillgen/expected/graphify__skill-vscode.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -307,7 +307,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -441,6 +441,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill-windows.md b/tools/skillgen/expected/graphify__skill-windows.md
index 2852de234..2642d9f5a 100644
--- a/tools/skillgen/expected/graphify__skill-windows.md
+++ b/tools/skillgen/expected/graphify__skill-windows.md
@@ -201,7 +201,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -246,7 +246,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -333,7 +333,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -467,6 +467,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/expected/graphify__skill.md b/tools/skillgen/expected/graphify__skill.md
index 6c7060a55..29c5f66a5 100644
--- a/tools/skillgen/expected/graphify__skill.md
+++ b/tools/skillgen/expected/graphify__skill.md
@@ -179,7 +179,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -224,7 +224,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -311,7 +311,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -445,6 +445,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").
diff --git a/tools/skillgen/fragments/core/core.md b/tools/skillgen/fragments/core/core.md
index 12b8a16a6..5d914de25 100644
--- a/tools/skillgen/fragments/core/core.md
+++ b/tools/skillgen/fragments/core/core.md
@@ -138,7 +138,7 @@ for f in detect.get('files', {}).get('code', []):
     code_files.extend(collect_files(Path(f)) if Path(f).is_dir() else [Path(f)])
 
 if code_files:
-    result = extract(code_files, cache_root=Path('.'))
+    result = extract(code_files, cache_root=Path('INPUT_PATH'))
     Path('graphify-out/.graphify_ast.json').write_text(json.dumps(result, indent=2, ensure_ascii=False), encoding=\"utf-8\")
     print(f'AST: {len(result[\"nodes\"])} nodes, {len(result[\"edges\"])} edges')
 else:
@@ -183,7 +183,7 @@ detect = json.loads(Path('graphify-out/.graphify_detect.json').read_text(encodin
 # every source file (#1392). Video is transcribed to a document in Step 2.5 first.
 all_files = [f for cat in ('document', 'paper', 'image') for f in detect['files'].get(cat, [])]
 
-cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files)
+cached_nodes, cached_edges, cached_hyperedges, uncached = check_semantic_cache(all_files, root='INPUT_PATH')
 
 # Always (re)write the cache file: write hits, else DELETE any leftover from a prior
 # run so Part C never merges a stale .graphify_cached.json (#1392).
@@ -246,7 +246,7 @@ from graphify.cache import save_semantic_cache
 from pathlib import Path
 
 new = json.loads(Path('graphify-out/.graphify_semantic_new.json').read_text(encoding=\"utf-8\")) if Path('graphify-out/.graphify_semantic_new.json').exists() else {'nodes':[],'edges':[],'hyperedges':[]}
-saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []))
+saved = save_semantic_cache(new.get('nodes', []), new.get('edges', []), new.get('hyperedges', []), root='INPUT_PATH')
 print(f'Cached {saved} files')
 "
 ```
@@ -380,6 +380,32 @@ If this step prints `ERROR: Graph is empty`, stop and tell the user what happene
 
 Replace INPUT_PATH with the actual path.
 
+### Step 4.5 - Graph health check (read-only integrity gate)
+
+A non-destructive diagnostic on the extraction, before labeling. It surfaces edge collapse, dangling/missing endpoints, and self-loops — the silent-corruption modes of incremental updates and AST/LLM id mismatches. Read-only; never aborts.
+
+```bash
+$(cat graphify-out/.graphify_python) -c "
+import json
+from pathlib import Path
+from graphify.diagnostics import diagnose_extraction, format_diagnostic_report
+
+extraction = json.loads(Path('graphify-out/.graphify_extract.json').read_text(encoding=\"utf-8\"))
+summary = diagnose_extraction(extraction, directed=IS_DIRECTED, root='INPUT_PATH')
+print(format_diagnostic_report(summary))
+flags = [f'{summary[k]} {label}' for k, label in (
+    ('dangling_endpoint_edges', 'dangling-endpoint edges'),
+    ('missing_endpoint_edges', 'missing-endpoint edges'),
+    ('self_loop_edges', 'self-loop edges'),
+    ('directed_same_endpoint_collapsed_edges', 'collapsed (directed) edges'),
+    ('undirected_same_endpoint_collapsed_edges', 'collapsed (undirected) edges'),
+) if summary.get(k, 0)]
+print('GRAPH HEALTH WARNING: ' + '; '.join(flags) + ' - graph may be incomplete/corrupt.' if flags else 'Graph health: OK (no dangling/missing/collapsed edges).')
+"
+```
+
+Substitute `IS_DIRECTED` and `INPUT_PATH` as in Step 4. If a `GRAPH HEALTH WARNING` prints, surface it in the final summary (do not abort — the graph is still usable, but the integrity issue must be visible, per the Honesty Rules).
+
 ### Step 5 - Label communities
 
 Read `graphify-out/.graphify_analysis.json`. For each community key, look at its node labels and write a 2-5 word plain-language name (e.g. "Attention Mechanism", "Training Pipeline", "Data Loading").