OpenClaw Docs Local Search Skill
Purpose
Use local DOCS/ corpus first for fast keyword lookup, then fall back to direct web fetch when needed.
Canonical root:
C:\Users\Miller\projects\clawdocs
Corpus path:
C:\Users\Miller\projects\clawdocs\DOCS
Prerequisite Check
Run:
bash
rg --version
If missing:
bash
brew install ripgrep
Query Routing
- •Broad topic lookup (file discovery):
bash
rg -il "<terms>" "C:\Users\Miller\projects\clawdocs\DOCS" --type md
- •Specific content search (context lines):
bash
rg -C 3 "<terms>" "C:\Users\Miller\projects\clawdocs\DOCS" --type md
- •Known file read:
bash
cat "C:\Users\Miller\projects\clawdocs\DOCS\<path>"
Quick Reference Index
- •
automation/ - •
channels/ - •
cli/ - •
concepts/ - •
experiments/ - •
gateway/ - •
help/ - •
install/ - •
platforms/ - •
reference/ - •
security/ - •
start/ - •
tools/ - •
web/
Update this list after each crawl/fetch cycle based on top-level folders in DOCS/.
Fallback
If local search returns no results, fetch directly from docs site:
https://docs.openclaw.ai/<path>
Use web_fetch against that URL.