Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →nvk/wiki-query
Fast, read-only, index-first queries over an llm-wiki with bounded file reads, exact citations, and honest evidence gaps.
Passed 3 of 14 test cases with the skill, and 2 of 14 without it.
and 6 more in the full report
5 fixed0 passed either way14 still failing3 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 4 test cases the skill fixed. Same model, same prompt, with and without the skill.
Query the system for 'auth'. We have 'auth-v1', 'auth-v2', 'auth-service', and 'authoring' topics in the hub index.
| Checked | Without skill | With skill |
|---|---|---|
| Presents no more than three candidate topics derived from the index | Missed | Passed |
| Asks a single clarifying question rather than scanning multiple topic directories | Passed | Passed |
The instructions your model reads when the skill loads.
Use this protocol for fast, read-only questions and inventory lookups. It is the canonical query profile shared by Claude, Codex, Pi, local models, and the portable fallback.
node_modules,
or every sibling topic.--local, or the current project contains .wiki/,
use <cwd>/.wiki and read .wiki/_index.md first.~/.config/llm-wiki/config.json. Expand only a leading ~
in hub_path. If unavailable, try resolved_path, then ~/wiki.<hub>/_index.md and <hub>/wikis.json. Choose exactly one
active topic from its title, aliases, summary, or an explicit --wiki NAME.
Resolve registry paths relative to the hub; if stale, try
<hub>/topics/NAME._index.md, then only the relevant branch
index: wiki/_index.md, raw/_index.md, inventory/_index.md,
datasets/_index.md, or output/_index.md.If topic choice is genuinely ambiguous, list at most three index-derived candidates and ask one short question instead of scanning multiple topics.
test cases passed, vs 2 / 14 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.