explore

Name: explore
Author: xwtro0tk1t-cloud

320stars42forks

How to install explore?

xwtro0tk1t-cloud/harness/explore

$npx -y skills add xwtro0tk1t-cloud/harness --skill explore

Installs into the current project.

Summary

Graph-driven project understanding using code-review-graph (CRG). Query architecture, modules, callers/callees, impact radius, hotspots, execution flows, and search nodes. Use when: (1) brainstorming and need to understand project structure, (2) writing plans and need impact anal

Files · 1

View on GitHub

SKILL.md

1	# Explore — Graph-Driven Project Understanding
2
3	## Overview
4
5	Query the project's code knowledge graph to understand architecture, trace call chains, and analyze impact. Powered by CRG's graph.db (SQLite) + CRG Python API.
6
7	Announce at start: "I'm using the explore skill to query the project's code knowledge graph."
8
9	## Prerequisites
10
11	- CRG must be installed: `python -c "import code_review_graph.tools"`
12	- Graph must exist: `.code-review-graph/graph.db`
13	- If missing → tell user: "Please run `/graph build` first"
14	- If graph older than 24h → suggest: "Graph may be stale, consider `/graph update`"
15
16	## Commands
17
18	### `/explore architecture` — Project Architecture Overview
19
20	Shows module decomposition, coupling relationships, and entry points.
21
22	```python
23	from pathlib import Path
24	from code_review_graph.tools import list_communities_func, get_architecture_overview_func
25
26	# Get communities (modules) — CRG takes repo_root, not db_path
27	repo = Path(".")
28	result = list_communities_func(repo_root=repo)
29	communities = result.get("communities", [])
30
31	# Get architecture overview
32	overview = get_architecture_overview_func(repo_root=repo)
33	```
34
35	Present as: module list with node counts, inter-module coupling, key entry points.
36
37	### `/explore module <name>` — Module Details
38
39	Shows nodes, edges, and responsibilities of a specific module.
40
41	```python
42	from pathlib import Path
43	from code_review_graph.tools import list_communities_func
44
45	# CRG has no single-community query API — filter from full list
46	result = list_communities_func(repo_root=Path("."))
47	communities = result.get("communities", [])
48	target = [c for c in communities if name.lower() in str(c).lower()]
49	```
50
51	### `/explore callers <func>` — Who Calls This Function?
52
53	Direct SQLite query (no CRG wrapper available for this):
54
55	```python
56	import sqlite3
57	conn = sqlite3.connect(".code-review-graph/graph.db")
58	conn.row_factory = sqlite3.Row
59	rows = conn.execute(
60	"SELECT * FROM edges WHERE target_qualified LIKE ? AND kind='CALLS'",
61	(f"%{func}%",)
62	).fetchall()
63	conn.close()
64	```
65
66	### `/explore callees <func>` — What Does This Function Call?
67
68	```python
69	import sqlite3
70	conn = sqlite3.connect(".code-review-graph/graph.db")
71	conn.row_factory = sqlite3.Row
72	rows = conn.execute(
73	"SELECT * FROM edges WHERE source_qualified LIKE ? AND kind='CALLS'",
74	(f"%{func}%",)
75	).fetchall()
76	conn.close()
77	```
78
79	### `/explore impact <file>` — Impact Analysis
80
81	Shows which flows and functions are affected by changes to a file.
82
83	```python
84	from pathlib import Path
85	from code_review_graph.tools import get_impact_radius, get_affected_flows_func
86
87	# CRG takes changed_files (list) and repo_root, not file_path + db_path
88	repo = Path(".")
89	impact = get_impact_radius(changed_files=["<file>"], repo_root=repo)
90	result = get_affected_flows_func(changed_files=["<file>"], repo_root=repo)
91	flows = result.get("flows", result.get("affected_flows", []))
92	```
93
94	### `/explore hotspots` — Hub Nodes, Bridge Nodes, Knowledge Gaps
95
96	IMPORTANT: `get_hub_nodes_func` / `get_bridge_nodes_func` from `code_review_graph.tools` are buggy. Use `code_review_graph.analysis` instead:
97
98	```python
99	from code_review_graph.tools._common import _get_store
100	from code_review_graph.analysis import find_hub_nodes, find_bridge_nodes
101
102	# CRG analysis functions take a store object, not db_path
103	# Must use _get_store() first to get the store
104	store, _ = _get_store(str(Path(".").resolve()))
105	hubs = find_hub_nodes(store, top_n=20)
106	bridges = find_bridge_nodes(store, top_n=20)
107	```
108
109	For knowledge gaps (`get_knowledge_gaps` has sqlite3.Row type issue), use direct SQLite:
110
111	```python
112	import sqlite3
113	conn = sqlite3.connect(".code-review-graph/graph.db")
114	# Nodes with high connectivity but no docstring/comments
115	rows = conn.execute("""
116	SELECT n.qualified_name, n.kind, COUNT(e.id) as edge_count
117	FROM nodes n
118	JOIN edges e ON n.qualified_name = e.source_qualified OR n.qualified_name = e.target_qualified
119	WHERE n.docstring IS NULL OR n.docstring = ''
120	GROUP BY n.qualified_name
121	HAVING edge_count > 5
122	ORDER BY edge_count DESC
123	LIMIT 20
124	""").fetchall()
125	conn.close()
126	```
127
128	### `/explore flows` — Execution Flow List
129
130	```python
131	from pathlib import Path
132	from code_review_graph.tools import list_flows
133
134	# Note: NO _func suffix — it's list_flows, not list_flows_func
135	result = list_flows(repo_root=Path("."))
136	flows = result.get("flows", [])
137	```
138
139	### `/explore search <keyword>` — Search Nodes by Keyword
140
141	Try CRG's search first, f

People who liked this also viewed

6 picks

Type

explore

320stars42forks

View on GitHub

How to install explore?

xwtro0tk1t-cloud/harness/explore

$npx -y skills add xwtro0tk1t-cloud/harness --skill explore

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Explore — Graph-Driven Project Understanding
2
3	## Overview
4
5	Query the project's code knowledge graph to understand architecture, trace call chains, and analyze impact. Powered by CRG's graph.db (SQLite) + CRG Python API.
6
7	Announce at start: "I'm using the explore skill to query the project's code knowledge graph."
8
9	## Prerequisites
10
11	- CRG must be installed: `python -c "import code_review_graph.tools"`
12	- Graph must exist: `.code-review-graph/graph.db`
13	- If missing → tell user: "Please run `/graph build` first"
14	- If graph older than 24h → suggest: "Graph may be stale, consider `/graph update`"
15
16	## Commands
17
18	### `/explore architecture` — Project Architecture Overview
19
20	Shows module decomposition, coupling relationships, and entry points.
21
22	```python
23	from pathlib import Path
24	from code_review_graph.tools import list_communities_func, get_architecture_overview_func
25
26	# Get communities (modules) — CRG takes repo_root, not db_path
27	repo = Path(".")
28	result = list_communities_func(repo_root=repo)
29	communities = result.get("communities", [])
30
31	# Get architecture overview
32	overview = get_architecture_overview_func(repo_root=repo)
33	```
34
35	Present as: module list with node counts, inter-module coupling, key entry points.
36
37	### `/explore module <name>` — Module Details
38
39	Shows nodes, edges, and responsibilities of a specific module.
40
41	```python
42	from pathlib import Path
43	from code_review_graph.tools import list_communities_func
44
45	# CRG has no single-community query API — filter from full list
46	result = list_communities_func(repo_root=Path("."))
47	communities = result.get("communities", [])
48	target = [c for c in communities if name.lower() in str(c).lower()]
49	```
50
51	### `/explore callers <func>` — Who Calls This Function?
52
53	Direct SQLite query (no CRG wrapper available for this):
54
55	```python
56	import sqlite3
57	conn = sqlite3.connect(".code-review-graph/graph.db")
58	conn.row_factory = sqlite3.Row
59	rows = conn.execute(
60	"SELECT * FROM edges WHERE target_qualified LIKE ? AND kind='CALLS'",
61	(f"%{func}%",)
62	).fetchall()
63	conn.close()
64	```
65
66	### `/explore callees <func>` — What Does This Function Call?
67
68	```python
69	import sqlite3
70	conn = sqlite3.connect(".code-review-graph/graph.db")
71	conn.row_factory = sqlite3.Row
72	rows = conn.execute(
73	"SELECT * FROM edges WHERE source_qualified LIKE ? AND kind='CALLS'",
74	(f"%{func}%",)
75	).fetchall()
76	conn.close()
77	```
78
79	### `/explore impact <file>` — Impact Analysis
80
81	Shows which flows and functions are affected by changes to a file.
82
83	```python
84	from pathlib import Path
85	from code_review_graph.tools import get_impact_radius, get_affected_flows_func
86
87	# CRG takes changed_files (list) and repo_root, not file_path + db_path
88	repo = Path(".")
89	impact = get_impact_radius(changed_files=["<file>"], repo_root=repo)
90	result = get_affected_flows_func(changed_files=["<file>"], repo_root=repo)
91	flows = result.get("flows", result.get("affected_flows", []))
92	```
93
94	### `/explore hotspots` — Hub Nodes, Bridge Nodes, Knowledge Gaps
95
96	IMPORTANT: `get_hub_nodes_func` / `get_bridge_nodes_func` from `code_review_graph.tools` are buggy. Use `code_review_graph.analysis` instead:
97
98	```python
99	from code_review_graph.tools._common import _get_store
100	from code_review_graph.analysis import find_hub_nodes, find_bridge_nodes
101
102	# CRG analysis functions take a store object, not db_path
103	# Must use _get_store() first to get the store
104	store, _ = _get_store(str(Path(".").resolve()))
105	hubs = find_hub_nodes(store, top_n=20)
106	bridges = find_bridge_nodes(store, top_n=20)
107	```
108
109	For knowledge gaps (`get_knowledge_gaps` has sqlite3.Row type issue), use direct SQLite:
110
111	```python
112	import sqlite3
113	conn = sqlite3.connect(".code-review-graph/graph.db")
114	# Nodes with high connectivity but no docstring/comments
115	rows = conn.execute("""
116	SELECT n.qualified_name, n.kind, COUNT(e.id) as edge_count
117	FROM nodes n
118	JOIN edges e ON n.qualified_name = e.source_qualified OR n.qualified_name = e.target_qualified
119	WHERE n.docstring IS NULL OR n.docstring = ''
120	GROUP BY n.qualified_name
121	HAVING edge_count > 5
122	ORDER BY edge_count DESC
123	LIMIT 20
124	""").fetchall()
125	conn.close()
126	```
127
128	### `/explore flows` — Execution Flow List
129
130	```python
131	from pathlib import Path
132	from code_review_graph.tools import list_flows
133
134	# Note: NO _func suffix — it's list_flows, not list_flows_func
135	result = list_flows(repo_root=Path("."))
136	flows = result.get("flows", [])
137	```
138
139	### `/explore search <keyword>` — Search Nodes by Keyword
140
141	Try CRG's search first, f

People who liked this also viewed

6 picks

Type