graphify-agents
/graphify-agents
Wykorzystuje graf wiedzy graphify, by dawać agentom kontekst kodu przy małym zużyciu tokenów podczas researchu, dekompozycji zadań i analizy wpływu zmian.
- Autor
- vinnie357 ↗
- Repozytorium
- claude-skills ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 07.10.2026
Ocena i bezpieczeństwo
- Pliki w paczce
- 1
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Brak sygnałów. Repozytorium zmienione 07.10.2026, bez odwołań do wycofanych modeli i API.
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Wykorzystuje graf wiedzy graphify, by dawać agentom kontekst kodu przy małym zużyciu tokenów podczas researchu, dekompozycji zadań i analizy wpływu zmian.
Oryginalny opis (EN)
Use a graphify knowledge graph to give agents low-token codebase context — query the graph instead of reading raw files during research, decomposition, and impact analysis. Use when wiring graphify into an agent-loop, an Explore/research subagent, or an MCP-enabled agent, or when deciding when a graph query beats grepping and reading files.
Jak działa
- The Core Pattern: Query, Don't Read
- Integration with /core:agent-loop
- Keeping the Graph Fresh
- Exposing the Graph via MCP
- Registering graphify with Claude Code
- When NOT to Use the Graph
- Anti-Fabrication Requirements
Podgląd SKILL.md
---
name: graphify-agents
description: Use a graphify knowledge graph to give agents low-token codebase context — query the graph instead of reading raw files during research, decomposition, and impact analysis. Use when wiring graphify into an agent-loop, an Explore/research subagent, or an MCP-enabled agent, or when deciding when a graph query beats grepping and reading files.
license: MIT
---
# Graphify for Agents
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.
## The Core Pattern: Query, Don't Read
| 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.
## Integration with /core:agent-loop
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:
```bash
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.
## Keeping the Graph Fresh
A stale graph misleads agents the way stale docs do. Keep it current:
```bash
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).
## Exposing the Graph via MCP
`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 `/extras-claude-code:claude-agents` (sibling plugin, install extras-claude-code@vinnie357; MCP-enabled agent pattern) for declaring the server in an agent's tool set.
## Registering graphify with Claude Code
`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`.
## When NOT to Use the Graph
- Small repos (the benchmark shows ≈1x on a 6-file library) — grep+read is simpler.
- Questions about the exact current contents of one known file — read it.
- Anything where the graph has not been rebuilt since the relevant code changed — refresh first or read directly.
## Anti-Fabrication Requirements
- Run `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.
- Verify a graph answer against the actual source file before an agent acts on it.
- Confirm graph freshness (`graphify check-update`) before trusting a query in a long session.
- State which commands require a full clustered build vs a raw AST extraction (`affected` and community labels need clustering).
Wersje
W bazie nie ma innych skilli o nazwie graphify-agents.
Migawka bazy z 13.11.2026 · wersja paczki c3cd6e00c61f (skrót SHA-256 całej paczki)
Pobrania w ostatnich 30 dniach: 0 · ranking
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 100graphify
Zamienia folder z kodem, dokumentami, plikami PDF i obrazami w graf wiedzy, który agenci odpytują z wiersza poleceń zamiast wielokrotnie czytać pliki. - 100graphify-codebase
Wdraża i przeszukuje graf kodu graphify przy projektowaniu, przeglądzie i ustalaniu zasięgu wpływu zmian w bazie kodu. - 100graphify
Buduje i odpytuje graf wiedzy o kodzie i dokumentach przez CLI graphify, instaluje go przez mise i eksportuje do GraphML, SVG, Neo4j lub Obsidiana. - 100graphify-usage
Wyjaśnia, jak odpytywać graf wiedzy graphify o strukturę kodu, kiedy wybrać narzędzia grafowe zamiast grep i jak czytać wyniki zapytań. - 90project-knowledge-graph
Buduje graf wiedzy z kodu, dokumentów, artykułów i zrzutów ekranu, żeby szybko poznać strukturę dużego projektu i jego wiedzę. - 100code-knowledge-graph-tools-for-ai-agents
Porównuje GitNexus, Graphify i CodeGraph, narzędzia budujące graf wiedzy o kodzie dla agentów AI przez MCP, zastępując powtarzane przeszukiwanie grep zapytaniami o wpływ, łańcuchy wywołań i zależności.