Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →codealive-ai/exa-web-search
General-purpose web search powered by Exa — the search_web / web_search analog.
Passed 3 of 10 test cases with the skill, and 2 of 10 without it.
and 5 more in the full report
6 fixed2 passed either way14 still failing2 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.
I have the URL https://techblog.example.com/raft-consensus-impl and want to fetch and read its full page body text rather than just a summary snippet. Which CLI command should be run?
| Checked | Without skill | With skill |
|---|---|---|
| Invokes scripts/exa_client.py using the contents subcommand. | Missed | Passed |
| Includes the --text flag to retrieve full page text. | Missed | Passed |
The instructions your model reads when the skill loads.
Plain semantic web search — the everyday search_web you reach for first. No category, no domain restriction: describe what you're looking for and Exa returns the closest pages. Calls the Exa REST API through a local script; no MCP server needed, only an EXA_API_KEY.
This skill is self-contained — its client (scripts/exa_client.py) is vendored, so you can install just this folder.
Get a key first at https://dashboard.exa.ai/api-keys, then:
export EXA_API_KEY=your-key # add to ~/.zshrc / ~/.bashrc to make it permanent
Just this skill (Claude Code):
git clone --depth 1 https://github.com/CodeAlive-AI/exa-skills /tmp/exa-skills
cp -r /tmp/exa-skills/exa-web-search ~/.claude/skills/exa-web-search # user-level
# …or .claude/skills/exa-web-search inside a project for project-level
That's the whole install — no other folders required. (Other agents: drop the exa-web-search folder wherever your agent discovers skills.)
Whole collection: clone the repo and point your agent at it — see the repo README.
python scripts/web_search.py "<what you're looking for>" [-n N] [--text] [--json]
Examples:
python scripts/web_search.py "how do vector databases handle deletes" -n 10
python scripts/web_search.py "best practices for prompt caching with LLMs" --text
python scripts/web_search.py "self-hosted analytics alternatives to Google Analytics" -n 12
Output is a Markdown list (title · date · snippet) ending in sources_reviewed: N. --text includes page content, --json gives raw results, -h lists all flags.
test cases passed, vs 2 / 10 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.