MCP App Store
by KeenEthics
Sken.io icon

Sken.io MCP Server for ChatGPT

By Sken.io

Overview

Sken.io watches web pages for you and tells you when they change. Describe a page and what matters on it in plain language, and Sken.io starts monitoring it: the price on a product listing, a competitor's pricing table, a careers page, a booking calendar, a supplier's terms. You pick how often it looks, from once a minute to once a day. Every check is a real browser loading the real page, so Sken.io keeps a screenshot of the page before and after each change, with the changed area highlighted. That means you can ask what changed and get an actual answer about the page, instead of a link telling you to go and look yourself. You can also review recent changes across everything you monitor, pause or retarget a monitor, group monitors by project, and send alerts on to e-mail, Slack, Discord, Google Chat, webhooks or Zapier so your team hears about a change even when you are not in a chat. Requires a Sken.io account. Monitoring uses the check allowance on that account, and the connector reports how much of it is left when you ask.

How to connect

  1. ChatGPT: Open the Sken.io listing in the ChatGPT app directory and choose Connect. ChatGPT listing Sken.io exposes 43 tools here.

Tools

sken_account_info

ChatGPT
Current plan, remaining points, trial end date and job counts. Check this when an operation fails for quota reasons, or when the user asks about their plan or usage.

sken_create_export

ChatGPT
Queue a data export. Processing is asynchronous - the response confirms the export was queued, not that the file is ready. Rate limited more tightly than other endpoints, so do not poll it in a loop.

sken_create_group

ChatGPT
Create a group for organising jobs. A group is a label used to filter jobs and report on a set of pages together; it does not change how any job is checked. sken_list_groups shows the groups that already exist.

sken_create_import

ChatGPT
Create many monitoring jobs at once from a CSV passed in file. Processing is asynchronous - the response confirms the import was queued, not that the jobs exist; poll sken_list_imports for that. Leave deleteMissingJobs off unless the user explicitly asked for the account to mirror the file, because it deletes every job the file does not contain. Each imported job consumes account points.

sken_create_integration

ChatGPT
Add a notification channel. Get the type id from sken_list_integration_types first, then verify the result with sken_test_integration.

sken_create_job

ChatGPT
Start monitoring a URL. areaType picks what is watched: ai for a natural-language prompt evaluated by a vision model, element for XPaths, crop for a pixel rectangle. Creating the job and every check it then runs consume account points. areaType is ai unless the user named element or crop, or supplied a concrete XPath or pixel rectangle. An ai prompt takes a plain-language description of the target and needs no selector. An element XPath that matches nothing produces a job that reports no changes, which looks identical to a page that never changes. For element, each XPath has to take one of the two forms the Sken.io element picker produces: - /html/body - the whole page. - //*[@id="..."] - anchored on an element id, e.g. //*[@id="price"][1]. Other forms, such as //header//div[@class="header-icons"]/ul, are not supported. When the target cannot be written in one of those two forms, ask the user for an element id or use ai with a plain-language prompt. State the chosen area type and what it will catch in the reply.

sken_create_job_setting

ChatGPT
Create a settings profile several jobs can share. Use when the user wants the same cadence, proxy or notification wiring applied to several pages.

sken_create_proxy_user_endpoint

ChatGPT
Register a proxy the user already has, so jobs can run through it. Needs a name and the exact endpoint string, which has to come from the user and cannot be constructed.

sken_delete_group

ChatGPT
Permanently delete a group. The jobs inside it keep running and are not deleted, they only lose this grouping, and the grouping cannot be restored through the API.

sken_delete_integration

ChatGPT
Remove a notification channel. Jobs wired to it will no longer notify through it, so check which jobs use it before deleting.

sken_delete_job

ChatGPT
Permanently delete a job, together with its check history and stored screenshots. This cannot be undone. sken_update_job with active: false pauses a job without deleting it.

sken_delete_job_setting

ChatGPT
Permanently delete a shared settings profile. Jobs that used it keep running but lose the schedule, proxy and notification wiring it provided. sken_list_job_settings reports how many jobs still reference it. This cannot be undone.

sken_delete_proxy_user_endpoint

ChatGPT
Remove an own proxy endpoint. Job settings pointing at it lose their proxy, so check jobSettingsCount from sken_list_proxy_user_endpoints first.

sken_get_import_report

ChatGPT
The report for one import run, listing which rows failed and why. Read it when sken_list_imports shows a non-zero error count.

sken_get_job

ChatGPT
Full configuration of a single job: watched areas, schedule, proxy, page function, cookies and notification wiring. Use this before updating a job, so the change is made against its real current configuration.

sken_get_job_action_types

ChatGPT
The action types a job or a scan can run on the page before it is captured, such as an AI instruction that dismisses a cookie banner or clicks a tab, and a fixed wait. Returns each type key with its label, the operations it supports and which of them need a value. Call before putting anything in jobActions, because the type keys are fixed.

sken_list_changes

ChatGPT
Check results for this account, newest first. Use when the user asks what changed on a monitored page. Filter to results that detected a change with changeCode: 1, or to one page with jobIds. Each result carries a check id and screenshot paths; the images themselves are read as the resource sken://checks/{checkId}/screenshot, with /previous for the before-image and /marked for the diff-highlighted version. Each result also carries direct imgUrl and markedImgUrl links that can be passed to the user. This tool only reads; marking changes as seen is done with sken_mark_job_seen.

sken_list_export_types

ChatGPT
The export formats that can be requested, with the id each one needs. Call before sken_create_export.

sken_list_exports

ChatGPT
Past and pending exports with their status and file name. This is how you check whether an export queued by sken_create_export has finished.

sken_list_groups

ChatGPT
List the groups jobs can be organised into. Needed to resolve a group name to the id used by groupId filters and by sken_create_job.

sken_list_imports

ChatGPT
Bulk import runs with their status and row counts - processed, valid, failed. Use after sken_create_import to confirm the run finished and how many rows failed.

sken_list_integration_types

ChatGPT
The integration types that can be created, with the id each one needs. Call this before sken_create_integration - the type id is not guessable.

sken_list_integrations

ChatGPT
List configured notification channels (Slack, webhook, e-mail) and their ids, which jobs reference to decide where change alerts go.

sken_list_job_settings

ChatGPT
Reusable settings profiles - schedule, proxy, notification wiring, colour - that jobs share by id instead of repeating the configuration. Use to find the id of a profile before attaching it to a job.

sken_list_jobs

ChatGPT
List the monitoring jobs on this account. Use to resolve a page the user names or links to its job id, and to find jobs with unseen changes (hasNewChanges: true). Returns id, url, name, active flag, schedule and next check time.

sken_list_proxy_endpoints

ChatGPT
Proxy endpoints Sken.io provides, with country and provider. Their ids go into proxyEndpointId on a job or job setting, which makes checks run from that country. Relevant for pages that geo-block or localise their content.

sken_list_proxy_user_endpoints

ChatGPT
The user's own proxy endpoints and how many job settings use each. Their ids go into proxyUserEndpointId.

sken_mark_job_seen

ChatGPT
Clear the unseen-changes flag on a job, the same as opening it in the web app. Call this after reporting a job's changes to the user, so the same change is not reported again.

sken_proxies_overview

ChatGPT
Proxy subusers purchased on this account and their status. Use when the user asks which proxies the account owns.

sken_proxy_pricing

ChatGPT
The current price per gigabyte of Sken.io proxy traffic. Use when the user asks what proxy traffic costs.

sken_revert_job_element_heal

ChatGPT
Restore the element the user originally picked on a job, discarding the automatic re-location the engine performed after the stored XPath stopped matching. Use when an automatically re-located job is reporting changes on the wrong part of the page.

sken_scan_website

ChatGPT
Take an immediate one-off screenshot of a URL without creating a job. Use it to preview a page before setting up monitoring, or to check a page right now. Returns straight away with a wsRoom handle while the scan runs in the background, so the screenshot may not be ready in the response. The scan consumes account points.

sken_stats_by_job

ChatGPT
Checks, changes and errors broken down per job. Use to rank jobs - which page changes most, which one keeps erroring.

sken_stats_checks

ChatGPT
Checks, changes and errors as a time series over a date range. Use for trends and for spotting the day something started failing.

sken_stats_jobs

ChatGPT
How the account is set up: active vs inactive jobs, how many have notifications wired up, the split between fixed-frequency and weekly schedules, and how many jobs were created in a date range. This covers configuration; check activity totals are in sken_stats_overview.

sken_stats_overview

ChatGPT
Account-level totals: how many jobs exist, how many are active, and how many checks, changes and errors fell in a date range. Use when the user asks for a summary of monitoring activity over a period.

sken_test_integration

ChatGPT
Send a test message through a notification channel to confirm it is configured correctly. The message is delivered for real to the Slack workspace, Discord server, Google Chat space, webhook endpoint or e-mail address stored on the integration, and cannot be recalled once sent.

sken_trigger_checks

ChatGPT
Bring the next check forward to now. With jobIds, only those jobs run; with the field omitted, every active job on the account runs. Each triggered check loads the monitored page in a real browser and consumes account points, so omitting jobIds is expensive on an account with many jobs. The response gives the number of jobs triggered, not what they found - read that with sken_list_changes once the checks have run.

sken_update_group

ChatGPT
Rename a group or change the colour it is shown in. The jobs in the group are not affected. Resolve the id with sken_list_groups first.

sken_update_integration

ChatGPT
Change a notification channel - its name, webhook URL, Slack target or e-mail address. Confirm the result with sken_test_integration afterwards.

sken_update_job

ChatGPT
Change an existing job - pause or resume it (active), adjust the schedule, swap the proxy, retarget the watched area. Only the fields you send are changed. Retargeting the watched area follows the same rules as creating one: areaType is ai unless the user named element or crop, and any XPath takes the form /html/body or //*[@id="..."].

sken_update_job_setting

ChatGPT
Change a shared settings profile. The change applies to every job using it; sken_list_job_settings reports how many that is.

sken_update_proxy_user_endpoint

ChatGPT
Rename a proxy endpoint the user registered, or point it at a different address. Job settings using it switch to the new address on their next check.

Capabilities

Writes

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/sken-io"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/sken-io)