/graph
Compile, lint, or query the wiki's typed graph layer.
$ npx -y skills add praneybehl/llm-wiki-plugin --agent claude-codeHow it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/graph
Context preview
What this command does when you run it.
Compile, lint, or query the wiki's typed graph layer.
Command definition
graph.mddescription: Compile, lint, or query the wiki's typed graph layer.
argument-hint: "extract | lint | neighbors --node <id> | edges --subject <id> | path --from <id> --to <id> | facts --about <id>"
Run a graph-layer operation against the wiki using the `llm-wiki` skill's graph workflow.
Arguments: $ARGUMENTS
1. Confirm `wiki/graph/ontology.yaml` exists. If not, the wiki is pre-graph — offer to seed `wiki/graph/` (copy the templates from `skills/llm-wiki/assets/`) before doing anything else. Do not fabricate ontology entries. 2. Dispatch on the first argument:
- `extract` → `uv run --script "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_extract.py" "${PWD}/wiki"`
- `lint` → `uv run --script "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_lint.py" "${PWD}/wiki"`
- `neighbors --node <id>` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" neighbors --node <id>`
- `edges --subject <id> [--predicate <p>]` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" edges --subject <id> [--predicate <p>]`
- `path --from <id> --to <id> [--max-depth N]` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" path --from <id> --to <id>`
- `facts --about <id>` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" facts --about <id>`
3. For query subcommands, after running the script, follow the most relevant edges back to their wiki pages and summarize. Cite with `[[wikilinks]]` to wiki pages, not graph rows. The graph accelerates navigation; the wiki page and its raw source remain the evidence. 4. For `lint`, present findings as proposed edits — do not silently rewrite typed `graph.relationships`. After approved fixes, suggest re-running `extract` so the compiled artifacts stay in sync with the markdown. 5. If `graph.sqlite` is older than the most recent ingest line in `log.md`, suggest running `extract` first.
Full reference: `skills/llm-wiki/references/graph-workflow.md`.
Read more
description: Compile, lint, or query the wiki's typed graph layer. argument-hint: "extract | lint | neighbors --node <id> | edges --subject <id> | path --from <id> --to <id> | facts --about <id>"
Run a graph-layer operation against the wiki using the `llm-wiki` skill's graph workflow.
Arguments: $ARGUMENTS
1. Confirm `wiki/graph/ontology.yaml` exists. If not, the wiki is pre-graph — offer to seed `wiki/graph/` (copy the templates from `skills/llm-wiki/assets/`) before doing anything else. Do not fabricate ontology entries. 2. Dispatch on the first argument:
- `extract` → `uv run --script "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_extract.py" "${PWD}/wiki"`
- `lint` → `uv run --script "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_lint.py" "${PWD}/wiki"`
- `neighbors --node <id>` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" neighbors --node <id>`
- `edges --subject <id> [--predicate <p>]` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" edges --subject <id> [--predicate <p>]`
- `path --from <id> --to <id> [--max-depth N]` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" path --from <id> --to <id>`
- `facts --about <id>` → `python "${CLAUDE_PLUGIN_ROOT}/skills/llm-wiki/scripts/wiki_graph_query.py" "${PWD}/wiki" facts --about <id>`
3. For query subcommands, after running the script, follow the most relevant edges back to their wiki pages and summarize. Cite with `[[wikilinks]]` to wiki pages, not graph rows. The graph accelerates navigation; the wiki page and its raw source remain the evidence. 4. For `lint`, present findings as proposed edits — do not silently rewrite typed `graph.relationships`. After approved fixes, suggest re-running `extract` so the compiled artifacts stay in sync with the markdown. 5. If `graph.sqlite` is older than the most recent ingest line in `log.md`, suggest running `extract` first.
Full reference: `skills/llm-wiki/references/graph-workflow.md`.
Turn PDFs, articles, transcripts, and notes into a shared wiki that your AI agents can search, cite, and keep up to date. Add a source once. Ask questions later. Keep the useful answers.
Other commands on praneybehl-llm-wiki.
- /ingest
Ingest a source (paper, article, transcript, PDF, notes) into the LLM Wiki.
Open command - /init
Initialize a new LLM Wiki structure in this project (creates wiki/ and raw/ directories with templates).
Open command - /lint
Run a structural and semantic health check on the LLM Wiki.
Open command - /query
Query the LLM Wiki — answer a question from accumulated knowledge with citations.
Open command - /stats
Show wiki size, shape, and link density — and which scaling threshold the wiki is at.
Open command - /upgrade
Upgrade an existing LLM Wiki to the current plugin version (adds missing files idempotently; walks me through any SCHEMA.md merges by hand).
Open command