Help us improve
Share bugs, ideas, or general feedback.
From all-skills
Gives agents low-token codebase context by querying a graphify knowledge graph instead of reading raw files. Use for research, impact analysis, and agent-loop integration.
npx claudepluginhub vinnie357/claude-skills --plugin qaHow this skill is triggered — by the user, by Claude, or both
Slash command
/all-skills:graphify-agentsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Reading raw files into context is the dominant token cost of codebase-understanding work. Graphify builds a knowledge graph once (`graphify-out/graph.json`) and answers structural questions by traversing it — returning the relevant nodes and edges instead of whole files. This skill covers using that graph to make agent workflows cheaper and more targeted.
Builds a local knowledge graph of code, docs, and mixed media using tree-sitter extraction and LLM-based relationship inference. Query via CLI commands instead of reading raw files.
Uses codebase knowledge graphs for architecture-aware task decomposition, dependency discovery, and context reduction during planning.
Dispatches code-review-graph queries as an agent: ensures graph freshness/embeddings, runs semantic search, feature exploration, impact analysis, and git-based review context.
Share bugs, ideas, or general feedback.
Reading raw files into context is the dominant token cost of codebase-understanding work. Graphify builds a knowledge graph once (graphify-out/graph.json) and answers structural questions by traversing it — returning the relevant nodes and edges instead of whole files. This skill covers using that graph to make agent workflows cheaper and more targeted.
The project's published benchmark reports "71.5x fewer tokens per query vs reading raw files" on a mixed corpus, with smaller gains (≈5.4x on a 4-file corpus, ≈1x on a 6-file library) — the advantage scales with corpus size. Measure a specific repo with graphify benchmark before quoting a number; small repos see little benefit, so reserve the graph for large corpora.
For installing graphify and the full CLI surface, load the graphify skill.
Activate when:
/core:agent-loop Phase 1 pre-flight| Question shape | Without graphify | With graphify |
|---|---|---|
| "How does auth connect to the request pipeline?" | Grep, open 6–10 files, read each | graphify query "how does auth connect to the request pipeline?" → relevant nodes + edges |
"What is SwinTransformer and what touches it?" | Read the class + every caller | graphify explain "SwinTransformer" |
"What breaks if I change add?" | Trace callers by hand | graphify affected "add" |
| "How do these two modules connect?" | Read both, infer | graphify path "DigestAuth" "Response" |
query returns within a token budget (--budget, default 2000), so an agent gets a bounded, relevant slice rather than unbounded file contents.
Graphify slots into the loop's existing phases without replacing them. Load /core:agent-loop for the phase model.
Phase 1 (Pre-flight / research). Before decomposition, build or refresh the graph, then have the research/Explore subagent query it instead of fanning out file reads:
mise run graphify:update # AST-only refresh, no API key needed
graphify query "where is <epic-area> implemented and what does it depend on?"
Feed the query result into the Team Leader's decomposition. The graph names the real files and symbols to reference in worker prompts — which the agent-loop prompt template already requires ("reference existing code and functions to reuse").
Phase 2 (Working). Give each worker the path/explain output for its slice instead of pre-reading files into the prompt. Workers still read the specific files they edit.
Impact analysis before slicing. graphify affected "<symbol>" (needs a full clustered build) surfaces the blast radius of a change, which informs dependency edges between bees issues.
A stale graph misleads agents the way stale docs do. Keep it current:
graphify hook install # rebuild on post-commit / post-checkout
graphify watch . # rebuild live during a session
The agent that relies on the graph confirms freshness — graphify check-update . reports whether a semantic re-extraction is pending. Treat graph claims like any other: an agent verifies a graph answer against the actual file before acting on it (anti-fabrication).
graphify . --mcp starts an MCP stdio server (the [mcp] extra / graphify-mcp console script). An MCP-enabled agent then calls graphify tools directly rather than shelling out. See /claude-code:claude-agents (MCP-enabled agent pattern) for declaring the server in an agent's tool set.
graphify claude install writes a graphify section to CLAUDE.md and a PreToolUse hook so a Claude Code session reaches for the graph automatically. This is graphify's own integration; this marketplace's graphify/graphify-agents skills are the alternative, progressive-disclosure path that does not modify CLAUDE.md.
graphify benchmark (or cite the project's published figures as such) before stating a token-savings number — never present a benchmark figure as independent measurement.graphify check-update) before trusting a query in a long session.affected and community labels need clustering).