zyte-api-automation
by ComposioHQzyte-api-automation helps agents run Zyte API workflows through Composio Rube MCP by discovering current tool schemas, checking the zyte_api connection, and executing Web Scraping tasks with less guesswork.
This skill scores 67/100, which means it is acceptable for listing but should be presented as a lightweight integration guide rather than a full Zyte automation playbook. Directory users get enough evidence to know when to install it—when they use Rube MCP and need Zyte API access—but should expect to depend on live tool discovery for concrete operations and schemas.
- Valid frontmatter declares the required MCP dependency (`rube`) and the description clearly states the intended trigger: automating Zyte API tasks via Rube MCP.
- Setup prerequisites are explicit: connect Rube MCP, manage a `zyte_api` connection, verify ACTIVE status, and call `RUBE_SEARCH_TOOLS` before workflows.
- The workflow emphasizes schema discovery before execution, which should reduce stale-tool guesswork for agents using Composio/Rube.
- No support files, examples, scripts, or README are present beyond SKILL.md, so users must rely entirely on dynamic Rube tool discovery for exact schemas and execution details.
- The content is a generic MCP/toolkit wrapper pattern rather than a detailed Zyte-specific playbook, with limited practical examples or task-specific troubleshooting.
Overview of zyte-api-automation skill
What zyte-api-automation is for
zyte-api-automation is a Claude skill for running Zyte API workflows through Composio’s Rube MCP toolkit. It is designed for agents that need to discover current Zyte API tool schemas, verify the user’s Zyte connection, and then execute web scraping or extraction tasks with less guesswork than a plain prompt.
The practical value is not “scrape a page” by itself. The value is the workflow discipline: search the available Rube tools first, use the returned schema instead of assuming parameters, check the zyte_api connection, and only then run the task.
Best-fit users and tasks
This skill fits users who already use, or plan to use, Rube MCP with Composio and need Zyte API automation inside an AI-agent workflow. It is especially relevant for teams building assistants for Web Scraping, browser-rendered page retrieval, structured extraction, crawl support, or Zyte API task orchestration.
It is less useful if you only need a static scraping script, do not use MCP tools, or want a full crawler framework with scheduling, storage, retries, and monitoring included.
What makes the skill different
The main differentiator is that zyte-api-automation tells the agent to call RUBE_SEARCH_TOOLS before acting. That matters because Composio tool schemas can change, and guessing tool names or fields is a common source of failed automation.
The upstream skill is compact and contains no helper scripts, examples directory, or metadata file beyond SKILL.md. Install it for its MCP execution pattern, not for a large reference library.
How to Use zyte-api-automation skill
zyte-api-automation install context
Install the skill from the Composio skills repository:
npx skills add ComposioHQ/awesome-claude-skills --skill zyte-api-automation
Then configure Rube MCP in your client by adding https://rube.app/mcp as an MCP server. The skill expects RUBE_SEARCH_TOOLS to be available and requires an active Zyte API connection through RUBE_MANAGE_CONNECTIONS with toolkit zyte_api.
Before relying on the skill, open composio-skills/zyte-api-automation/SKILL.md. It is the primary source file and contains the required setup and workflow pattern.
Inputs the skill needs
A weak request is: “Use Zyte to scrape this site.”
A stronger prompt for zyte-api-automation usage is:
Use the
zyte-api-automationskill. First callRUBE_SEARCH_TOOLSfor current Zyte API operations and schemas. Check whether thezyte_apiconnection is ACTIVE. Then retrievehttps://example.com/productsand extract product name, price, availability, and canonical URL. Prefer browser rendering if the discovered Zyte tool supports it. Return the execution plan before running destructive or costly steps.
Good inputs include the target URL, desired output fields, rendering needs, crawl depth if any, authentication constraints, rate or cost sensitivity, and what format you want back.
Recommended workflow
For reliable Zyte API automation, follow this sequence:
- Ask the agent to use the skill by name.
- Have it call
RUBE_SEARCH_TOOLSwith your specific use case, not a generic phrase. - Check the
zyte_apiconnection withRUBE_MANAGE_CONNECTIONS. - If inactive, complete the returned authentication flow.
- Use the discovered tool slug and schema exactly.
- Review the planned request before execution when scraping volume, cost, or compliance risk matters.
This pattern is more important than memorizing a single Zyte tool name, because the skill itself emphasizes current schema discovery.
Practical prompt tips
Be explicit about success criteria. For example, say whether you need raw HTML, browser-rendered content, screenshots, structured JSON, or extracted fields. If you are scraping multiple pages, specify whether the agent should process a small sample first.
For zyte-api-automation for Web Scraping, also state boundaries: allowed domains, maximum pages, retry behavior, robots/compliance expectations, and whether personally identifiable information should be ignored or redacted.
zyte-api-automation skill FAQ
Is zyte-api-automation useful without Rube MCP?
No. The skill requires Rube MCP and depends on RUBE_SEARCH_TOOLS and RUBE_MANAGE_CONNECTIONS. If your AI client cannot call MCP tools, a normal prompt or direct Zyte API SDK integration will be more appropriate.
How is this better than an ordinary prompt?
An ordinary prompt may hallucinate Zyte parameters or call outdated tool names. The zyte-api-automation skill improves reliability by forcing tool discovery first, then connection validation, then execution using live schemas returned by Rube.
It does not guarantee scraping success. Site behavior, anti-bot protections, cost limits, authentication, and extraction ambiguity can still affect results.
Is it beginner-friendly?
It is beginner-friendly if you are comfortable connecting an MCP server and following an auth link. It is not a no-code scraping product. Beginners should start with one URL, one extraction goal, and a request to show the discovered tool schema before execution.
When should I not use this skill?
Do not use it when you need a complete crawling platform, long-running scheduler, database pipeline, or custom data-cleaning system. Also avoid it for sites you are not authorized to access or tasks where legal, privacy, or terms-of-service constraints are unresolved.
How to Improve zyte-api-automation skill
Improve zyte-api-automation results with better task framing
The biggest quality gain comes from replacing broad scraping goals with operational instructions. Include the target pages, extraction schema, acceptable output format, and fallback behavior.
Better:
Extract article title, author, publish date, main text, and final URL from these 10 URLs. Use
zyte-api-automation; discover tools first, verifyzyte_apiis ACTIVE, run one sample, show the parsed result, then continue only if the fields look correct.
This reduces wasted calls and makes it easier to catch schema or rendering issues early.
Common failure modes to watch
Common issues include inactive Zyte connections, skipped RUBE_SEARCH_TOOLS, stale assumed schemas, vague extraction fields, and oversized scraping batches. If a run fails, first ask the agent to show the discovered tool slug, required fields, connection status, and exact payload it planned to send.
If output is incomplete, clarify whether the page requires JavaScript rendering, pagination handling, location settings, cookies, or login context.
Iterate after the first output
Treat the first run as calibration. Start with one representative URL, inspect the returned content, then refine selectors, fields, rendering options, or pagination instructions. For larger jobs, ask the agent to produce a small sample table and a list of uncertainties before scaling.
This is especially important for Web Scraping tasks where page templates vary across categories, regions, or device views.
What the upstream skill could add
The current repository entry is useful but minimal. It would be stronger with example prompts for common Zyte API tasks, sample RUBE_SEARCH_TOOLS responses, troubleshooting notes for inactive connections, and before/after examples showing how a rough scraping request becomes a schema-aware Rube MCP workflow.