Stats
Actions
Tags
From exa-search
Guides effective use of built-in Exa tools for web search, code search, URL crawling, and deep research. Triggers when searching for external information, looking up docs/APIs, researching technologies, investigating topics, or crawling URLs.
How this skill is triggered — by the user, by Claude, or both
Slash command
/exa-search:exa-searchThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill when you need to search the web, find code examples, crawl a URL, investigate a topic, or conduct deep research.
Use this skill when you need to search the web, find code examples, crawl a URL, investigate a topic, or conduct deep research.
Trigger phrases: search for, look up, find examples, investigate about, research about, dive into, crawl URL, what is, how does X work
Invoke at the stable published path (a symlink the plugin's SessionStart hook refreshes every session):
"${CLAUDE_CONFIG_DIR:-$HOME/.claude}/exa-search/search.sh" <command> [options]
$CLAUDE_CONFIG_DIR IS exported into the Bash tool's subshell; $CLAUDE_PLUGIN_ROOT is NOT — it is only set for hook subprocesses. Using the plugin-root path from a Bash tool call expands to an empty string and runs /skills/.../search.sh: No such file. Always use the published path above.
Repo-checkout fallback (for tests/dev when the plugin is not installed): plugins/exa-search/skills/exa-search/scripts/search.sh.
search.sh <command> [options]
Commands:
search Search the web (default if no command given)
crawl Extract content from URLs via /contents endpoint
similar Find pages similar to a URL via /findSimilar endpoint
search.sh search -q "query" [options]
search.sh "query" # bare query works too
Options:
-q, --query Search query (required)
-n, --num-results Number of results (default: 10, max: 100)
-t, --type instant|fast|auto|deep-lite|deep|deep-reasoning (default: auto)
-c, --category company|research paper|news|personal site|financial report|people
--include-domains Comma-separated domains to include
--exclude-domains Comma-separated domains to exclude
--start-date Start published date (YYYY-MM-DD)
--end-date End published date (YYYY-MM-DD)
--include-text Text that must appear (single phrase, max 5 words)
--exclude-text Text to exclude (single phrase, max 5 words)
--highlights N Max highlight chars (default: 4000)
--with-text Include full text in results
--lean Strip image/favicon/subpages/entities — AI-prompt-ready output
search.sh crawl <url> [url...] [-m max_chars]
Options:
-m, --max-chars Max characters per page (default: 3000)
search.sh similar <url> [-n num_results] [--highlights N]
| Constraint | Detail |
|---|---|
--include-text / --exclude-text | Single phrase ONLY. Multiple values cause 400 errors. |
company/people categories | Reject --start-date, --end-date, --exclude-domains. Use other categories when those filters matter. |
people category | --include-domains only accepts supported profile domains (LinkedIn et al). |
deep-* types | Adds synthesis latency (~seconds). Use instant or fast for low-latency lookups. |
--with-text when you need full page content, not just highlights| jq '.results[] | {title, url}'--lean when feeding results back to an LLM — strips image/favicon/subpages/entities to keep token count lowProvides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
npx claudepluginhub falconiere/toolu --plugin exa-search