agentsclimarketplace

Wiki graph

Skill galaxyuliana/wiki-graph/skills/wiki-graph

Living knowledge graphs for markdown — one self-contained HTML file, auto-updated as you (or your agent) edit. Zero dependencies.

Install
npx -y skills add galaxyuliana/wiki-graph --skill wiki-graph

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 28 days oldThe repository was created 28 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Render a markdown wiki, docs repo, or notes vault as an interactive knowledge graph — an Obsidian-style force view (self-contained HTML) plus an agent-readable graph.json. Use whenever the user invokes /wiki-graph, asks to see, refresh, watch, or schedule the knowledge graph, wants the wiki or docs visualized, or asks about hubs, orphans, broken links, or clusters in a docs repo. Works on any folder of markdown files; no dependencies beyond Python 3.

SKILL.md

4.2 KB, ~1.0k tokens by cl100k_base, as published. Nobody here has run it

wiki-graph

Build the graph of the repo the user is working in and show it. Every node is a markdown file (frontmatter type/status/tags drive color, filters, and search); every edge is a real markdown link — both [text](path.md) and Obsidian [[wikilinks]].

All commands below use <skill-dir> for this skill's base directory (stated when the skill loads).

Default flow (manual build)

  1. ROOT = git rev-parse --show-toplevel (not a git repo → use the cwd).
  2. Pick the builder:
    • Repo vendors its own copy at <ROOT>/tools/wiki-graph/build.py: run that one — teams pin the builder in-repo for reproducibility.
    • Otherwise run the bundled copy: python3 "<skill-dir>/scripts/build.py" "$ROOT". Outputs land in $ROOT/.wiki-graph/ (override with a second positional arg if the user wants them elsewhere).
  3. Open wiki-graph.html from there (open on macOS, xdg-open on Linux, start on Windows) and relay the doc/link/orphan counts the script printed.
  4. Git stays clean automatically — .wiki-graph/ writes its own .gitignore. If the user wants the graph committed/shared, that's git add -f .wiki-graph.
  5. First build in a repo? Offer two one-time upgrades (do them only if the user agrees):
    • Append the knowledge-graph block to the repo's AGENTS.md (create the file if missing) so every coding agent — Codex, Cursor, Gemini CLI, Copilot, etc. — keeps the graph fresh and reads graph.json when asked about the docs. Template: integrations/AGENTS.md-snippet.md in this plugin's repo; adapt the builder path to this machine (use the vendored path if the repo has one, else this skill's scripts path).
    • Vendor the builder into the repo (cp -r <skill-dir>/scripts <ROOT>/tools/wiki-graph) if the team wants a version-pinned copy that works without this plugin installed.

Live mode (auto-updating)

  • In Claude Code with the plugin installed, a PostToolUse hook already rebuilds .wiki-graph/ after every markdown edit — building once (step above) is all it takes to turn that on for a repo.
  • When the user says live / watch / serve, run python3 "<skill-dir>/scripts/build.py" "$ROOT" --serve in the background, then tell them the URL it printed (default http://127.0.0.1:7177/wiki-graph.html). The open page polls graph.json and morphs in place — new docs flash into the web as they are written. Stop it when they ask (kill the background job).

Scheduled mode

When the user wants the graph rebuilt on a schedule (e.g. nightly):

  • Preferred, if your environment has a scheduler (Claude Code cron, CI cron): schedule python3 "<skill-dir>/scripts/build.py" "<ROOT>" --quiet.
  • Plain OS fallback — offer this crontab line (adjust paths): 0 7 * * * python3 <skill-dir>/scripts/build.py <ROOT> --quiet
  • For a public, always-fresh graph of a GitHub repo, point them at integrations/github-pages.yml in this plugin's repository — it publishes the graph to GitHub Pages on every push and on a daily cron.

Answering questions about the graph

Read $ROOT/.wiki-graph/graph.json (rebuild first if stale): nodes carry type, status, tags, in, out — in + out == 0 → orphan; high in → hub. edges are [from, to] doc paths. broken lists [source, target] links that point at missing files — offer to fix those when asked. --check exits 1 when broken links exist (useful in CI).

Notes

  • Optional per-repo config ROOT/.wiki-graph.json: {"title": "My Wiki", "roots": ["docs"], "ignore": ["vendor"]}.
  • Stdlib-only, ~1s on a 400-doc wiki; dot-dirs and node_modules are skipped.
  • Timeline replay appears automatically when git history (or frontmatter dates) span more than one day.

Gives 0 of the 12 instructions most docs writing skills give in ~1.0k tokens

Counted across 1,637 of the 3,044 authors here whose files we hold, read 2026-08-07

  • announce the skill at startin 54 of 1637, across 26 files
  • convert legacy doc files before editingin 45 of 1637, across 7 files
  • Predict questions readers might askin 42 of 1637, across 4 files
  • Generate clarifying questions for initial contextin 42 of 1637, across 3 files
  • Create document scaffold with placeholder textin 42 of 1637, across 3 files
  • Brainstorm content options for each sectionin 42 of 1637, across 3 files
  • Test document with fresh context-less instancein 42 of 1637, across 3 files
  • include exact file paths in every taskin 42 of 1637, across 15 files
  • ask interview questions one at a timein 42 of 1637, across 27 files
  • Apply surgical edits during refinementin 41 of 1637, across 2 files
  • Offer structured workflow or freeformin 40 of 1637, across 1 file
  • Ask for document meta-contextin 40 of 1637, across 2 files

Said here and by no other author read

  • Use git root as the working directory
  • Run the vendored builder if present
  • Run the bundled builder otherwise
  • Append the knowledge graph block to AGENTS.md on first build
  • Run the builder with the serve flag for live mode
  • Stop the background server when asked

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.