Appendix

Command reference

# per-repo build (inside Claude Code)
/graphify .                      # default
/graphify . --mode deep          # aggressive INFERRED edges from docs
/graphify . --update             # re-extract only changed files

# headless / CI
graphify extract . [--mode deep] [--backend gemini|claude|...]

# merge
graphify merge-graphs a/graphify-out/graph.json b/graphify-out/graph.json --out cross-repo-graph.json

# query a non-default graph
graphify query "<question>" --graph cross-repo-graph.json [--dfs] [--budget 1500]
graphify path  "<A>" "<B>"  --graph cross-repo-graph.json
graphify explain "<node>"   --graph cross-repo-graph.json

# visualize
graphify export html --graph cross-repo-graph.json   # writes graph.html next to the JSON

Regenerating the HTML with community names and full labels

scripts/regen_html.py does three things:

1. writes community_name onto every node of the merged JSON,
2. calls graphify's to_html() with the names,
3. gives every node a label font size, so labels appear as you zoom in instead of being hidden below graphify's 15%-of-max-degree threshold.

$(cat graphify-out/.graphify_python) scripts/regen_html.py cross-repo-graph.json cross-repo-graph.html

All 46 community names (merged graph)

id	Name	id	Name
0	Bid Review Agent Design	23	Airframe Vocabulary Tests
1	DB Core & Spec Doc Shape	24	Capability Explanation Grouping
2	Report Sparklines	25	Capability Explanation Phrasing
3	Document Harvesting	26	Airframe Fallback Gate
4	Spec Dictionary & Tech Reqs	27	Component Taxonomy Ingest
5	Classify & Report Generation	28	Whitespace Normalization Tests
6	Line Item Parsing & Status	29	OpenWebUI Pipeline Wrapper
7	DB Query Layer	30	Keyword & Taxonomy Seeding
8	Pydantic AI Orchestrator	31	Bid Facts Mining
9	Demand Fit Analysis	32	Buyer Type Classifier
10	DB Upserts & Ingest	33	OCR Recovery
11	MVP Bid Resolution	34	Boilerplate Quote Tests
12	Spec Completeness Skill	35	Bid Classification Feed
13	Bid Review MCP Server	36	Component Demand Query
14	Platform & Weight Classifiers	37	Harvested Bid Lookup
15	Capability Text Parsing	38	Recent Pipeline Runs
16	Keyword Bid Scraping	39	Modal Serving Concepts
17	LLM Pipeline Stage 4	40	Modal Smoke Tests
18	Line Item Use Case & Specs	41	GPU Serving Profiles
19	LLM Spec Mining	42	Modal Apps & Secrets
20	Pipeline Run Tracking	43	Qwen3 vLLM Servers
21	VTOL Category Audit	44	Laya T4 Server
22	Quote Verification Tests	45	modal-io-serve Package

Communities 0-38 are drone-okg and 39-45 are modal_io_serve. No community contains nodes from both repos.

Numbers at a glance

Graph	Nodes	Edges	Communities
drone-okg (default)	773	1,681	39
modal_io_serve (default)	64	86	7
Merged (default)	832	1,767	46
Merged + 2 manual calls_http edges	832	1,769	46
drone-okg (deep)	713	1,493	39
modal_io_serve (deep)	62	104	10
Merged (deep)	770	1,597	n/a