メニュー
Local knowledge graph for code and docs (Graphify). Extract a folder into graphify-out/graph.json, then query with CLI or MCP before bulk reads. Install: pip install 'graphifyy[mcp]'
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | graphify |
| description | Local knowledge graph for code and docs (Graphify). Extract a folder into graphify-out/graph.json, then query with CLI or MCP before bulk reads. Install: pip install 'graphifyy[mcp]' |
| requires.bins | graphify |
Turn a folder of code, schemas, docs, papers, or notes into a queryable NetworkX graph. Prefer graph queries over dumping whole trees into context.
Install: pip install 'graphifyy[mcp]'
Upstream: https://github.com/Graphify-Labs/graphify
read_file / grep sweeps when a graph already existsgraphify updateDo not paste graph.json or GRAPH_REPORT.md into the system prompt — query via CLI/MCP.
Prefer code-only extract. A plain extract without --code-only runs a
semantic/docs pass that needs an LLM backend; without keys it can exit 0 and
still leave no graph.json.
# Prefer ClawAgents layout (VS Code uses this path). Use python3 (not bare python).
export GRAPHIFY_OUT=.clawagents/graphify
python3 -m graphify extract . --code-only
# Augment after edits (AST, no LLM):
python3 -m graphify update .
Full extract (docs/images, needs LLM):
export GRAPHIFY_OUT=.clawagents/graphify
python3 -m graphify extract .
Upstream default output is graphify-out/ if GRAPHIFY_OUT is unset. If that
already exists, VS Code can Use existing graphify-out to copy it into
.clawagents/graphify/.
graphify query "authentication flow" --budget 2000
graphify path "AuthService" "UserStore"
graphify explain "ScopeGraph"
Or use MCP tools when the Graphify server is connected: query_graph, get_node, get_neighbors, shortest_path, god_nodes, graph_stats.
python3 -m graphify.serve .clawagents/graphify/graph.json
export GRAPHIFY_OUT=.clawagents/graphify
python3 -m graphify update .
graphify global add /path/to/notes --as notes
# serve ~/.graphify/global-graph.json via MCP, or set ClawAgents graphify_graph_path
~/.graphify/global-graph.jsongraphifyy[mcp] into the managed interpreter)| Path | Role |
|---|---|
.clawagents/graphify/graph.json | Canonical graph (ClawAgents default) |
graphify-out/graph.json | Upstream default |
GRAPH_REPORT.md | Human/agent summary (read on demand) |
wiki/ | Optional markdown wiki (--wiki) |
Implement a task N ways in parallel and pick the best. Spawns multiple subagents in isolated worktrees, evaluates candidates, and applies the winner. Use when asked to "best of n", "try multiple approaches", "parallel implementations", "/best-of-n", or "/bon".
Context database for AI agents. Use the `ov` CLI to store, organize, and retrieve project context (resources, memories, skills) in a hierarchical filesystem paradigm with tiered L0/L1/L2 loading. Query before work to load relevant context; add resources to build the knowledge base. Install: pip install openviking