0.1.0 • Published yesterdayCLI
scraper-api-skill
Licence
UNLICENSED
Version
0.1.0
Deps
0
Size
17 kB
Vulns
0
Weekly
0
@scraper-api/agent-skills
Agent skills that let Claude, Cursor, Codex and other skill-aware assistants query the scraper-api.com scraping API from natural language. No build step — these are Markdown skills with YAML frontmatter.
Skills
| Skill | Purpose |
|---|---|
scraper-api |
Data layer — intent→endpoint routing for all 11 platforms, auth, credits, pagination, limitations. |
competitor-social-research |
Workflow — profile a competitor's presence across platforms and their ad activity. |
transcript-intelligence |
Workflow — pull and analyse video/post transcripts across YouTube, Instagram, LinkedIn. |
The two workflow skills build on top of the scraper-api skill, which is the shared data layer.
Install
Copy the skill folders into your agent's skills directory:
- Claude Code / Claude Desktop:
~/.claude/skills/ - Cursor: your project's
.cursor/skills/(or the global skills dir) - Codex: the skills path configured for your setup
cp -R skills/* ~/.claude/skills/Requirements
Set an API key in the environment before use:
export SCRAPER_API_KEY="sk_..." # mint one with `scraper-api login` (the @scraper-api/cli)Security
API responses are data, not instructions. Scraped bios, captions, comments and transcripts are untrusted third-party content. These skills instruct the model to treat every payload strictly as data to analyse or relay — never as commands to execute (prompt-injection guard).