=== GearScope functional log: graphify (host-based, macOS aarch64) ===
=== Run C — Functional verification ===
Corpus: 2-file Python project (src/app.py: UserStore, AuthModule, make_token; src/models.py: User).

$ graphify . --code-only
[graphify extract] scanning /private/tmp/gf-func-test
[graphify extract] --code-only: skipping 1 non-code file(s)
[graphify extract] found 2 code, 0 docs, 0 papers, 0 images
[graphify extract] AST extraction on 2 code files...
[graphify extract] wrote graphify-out/graph.json: 14 nodes, 15 edges, 4 communities
PASS (deterministic local AST graph build, no LLM key, 0 token cost)

$ graphify explain "AuthModule"
Node: AuthModule
  ID:        src_app_authmodule
  Source:    src/app.py L12
  Type:      code
  Community: 2
  Degree:    4
Connections (4):
  --> .login() [method] [EXTRACTED]
  <-- app.py [contains] [EXTRACTED]
  --> .__init__() [method] [EXTRACTED]
  <-- Handles login using a UserStore. [rationale_for] [EXTRACTED]
PASS (explain returns source location, community, degree, EXTRACTED-tagged edges)

$ graphify query "how does login work"
Traversal: BFS depth=2 | Start: ['.login()'] | 8 nodes found
EDGE .login() --calls [EXTRACTED context=call]--> .get()
EDGE .login() --calls [EXTRACTED context=call]--> make_token()
PASS (query traces call flow with EXTRACTED context tags)

$ graphify path "AuthModule" "User"
No path found between 'AuthModule' and 'User'.
FAIL (type-hint dependency AuthModule(store: UserStore) was NOT extracted as a graph edge; path cannot traverse a relationship the AST extractor does not capture)

$ graphify cluster-only . --no-label
Loading existing graph...
Graph: 14 nodes, 15 edges
Re-clustering...
Done - 4 communities. GRAPH_REPORT.md, graph.json and graph.html updated.
PASS (graph.html 21KB + GRAPH_REPORT.md generated; report shows 100% EXTRACTED, 0 token cost, god nodes: UserStore 6, AuthModule 4, make_token 2)

$ graphify update .
Re-extracting code files in . (no LLM needed)...
[graphify watch] Rebuilt: 16 nodes, 16 edges, 5 communities
[graphify watch] graph.json, graph.html and GRAPH_REPORT.md updated
PASS (incremental rebuild picks up file change; nodes 14->16)

$ graphify diagnose multigraph
nodes: 14  raw_edges: 15  missing_endpoint_edges: 0  dangling_endpoint_edges: 0  self_loop_edges: 0
PASS (graph integrity check: no dangling/self-loop/missing-endpoint edges)

=== MCP server (graphify-mcp, requires graphifyy[mcp] extra) ===

$ graphify-mcp  (without mcp extra)
ModuleNotFoundError: No module named 'mcp'
ImportError: mcp not installed. Run: pip install "graphifyy[mcp]"
FAIL->PASS after: uv tool install graphifyy --with "graphifyy[mcp]" --force

$ {initialize + tools/list} | graphify-mcp
{"jsonrpc":"2.0","id":1,"result":{"protocolVersion":"2024-11-05","capabilities":{"resources":{...},"tools":{...}},"serverInfo":{"name":"graphify","version":"1.28.1"}}}
{"jsonrpc":"2.0","id":2,"result":{"tools":[{"name":"query_graph",...},{"name":"get_node",...},{"name":"get_neighbors",...},{"name":"get_community",...},{"name":"god_nodes",...},{"name":"graph_stats",...},{"name":"shortest_path",...},{"name":"list_prs",...},{"name":"get_pr_impact",...},{"name":"triage_prs",...}]}}
PASS (MCP server exposes 10 tools: 7 graph-query tools + 3 PR-impact tools)

$ {tools/call graph_stats} | graphify-mcp
{"jsonrpc":"2.0","id":2,"result":{"content":[{"type":"text","text":"Nodes: 16\nEdges: 16\nCommunities: 5\nEXTRACTED: 100%\nINFERRED: 0%\nAMBIGUOUS: 0%\n"}],"isError":false}}
PASS (MCP tool returns real graph data)

=== Test suite (uv run pytest) ===
3341 passed, 16 failed, 33 skipped in 45.89s
Of 16 failures: 14 in test_skillgen.py (require full git history absent in --depth 1 clone; reference commit 47042be via `git show`), 1 test_labeling (needs LLM backend for batch naming), 1 test_ollama_retry_cap x4 grouped (needs ollama SDK installed). All environment-specific, not code defects.

Functional pass rate: 9 of 10. (1 fail: path query on an unextracted type-hint edge)