llm-wiki: Command for Claude Code

Run the full llmwiki pipeline end-to-end: build → graph → export all → lint.

Usage: /wiki-all [flags]

$ARGUMENTS is forwarded verbatim to python3 -m llmwiki all. Common flags:

• --graph-engine builtin — skip optional Graphify (use when pip install llm-notebook[graph] has not been run)
• --skip-graph — skip the graph step entirely
• --strict — exit non-zero if lint reports any errors/warnings (good for CI)
• --fail-fast — stop at the first non-zero step instead of continuing to the next
• --with-synth — run synthesize before build (fills wiki/sources/ from raw/; opt-in because it may invoke an LLM)
• --synth-force — with --with-synth, pass --force to re-synthesize all sessions
• --out <dir> — output directory (default: site/)

Run:

python3 -m llmwiki all $ARGUMENTS

The command runs these steps in order and surfaces their combined output:

1. synthesize (only with --with-synth) — fill wiki/sources/ from raw/ via the configured LLM backend
2. build — compile the static HTML site from raw/ + wiki/
3. graph — build the knowledge graph (graph/graph.json + interactive graph.html)
4. export all — write every AI-consumable format (llms.txt, llms-full.txt, graph.jsonld, sitemap.xml, rss.xml, robots.txt, ai-readme.md)
5. lint — run every registered lint rule against the wiki

Report to the user:

• Output directory and total file / size count from the build step
• Graph stats (pages · edges · broken · orphans)
• Which export files were written
• Lint summary (errors / warnings / info)
• Overall exit code — 0 means every step succeeded

If any step fails, surface the failing step's output and the pipeline exit code.

Use this instead of chaining /wiki-build + /wiki-graph + /wiki-lint manually. It is the canonical one-shot "CI-ready site" command to run after /wiki-sync.