MCP Connector
The Skailer MCP server gives an AI assistant secure access to your HR data through OAuth authorization. Once connected, you can ask about employees, goals, time off, survey results, and performance reviews directly in a conversation - no exports and no API client required.
MCP (Model Context Protocol) is an open standard, so the server works with any client that supports it - Claude, other assistants, and your own integrations.
Connect
Section titled “Connect”If your client has a connector directory, add Skailer from there. For Claude:
https://claude.ai/directory/connectors/skailer
For any other MCP client, connect manually using the server address:
https://mcp.skailer.com/mcpOn first use the client redirects you to Skailer to sign in and approve the connection.
Choose what the assistant can do
Section titled “Choose what the assistant can do”The approval screen lists the permissions the application asks for:
- View data available to you - always on: no tool works without it.
- Create and update data available to you - you can uncheck it. The assistant will then only read data, and any attempt to change something is rejected.
A permission you unchecked is granted only when you connect the application again. Because assistants usually ask for both permissions, you will see the approval screen again on the next connection.
Whose data the assistant sees
Section titled “Whose data the assistant sees”The assistant acts on behalf of the person who approved the connection and sees exactly what that person would see in Skailer. It cannot read or change anything you have no access to. For example, a manager sees the time off of their team, and a review participant sees only the results shared with them.
What the server can do
Section titled “What the server can do”Profile and employees
Section titled “Profile and employees”get_me- your compact profile: name, position, department, manager, and current time off. The assistant uses it to learn who you are.search_employees- search employees by name, position, department, and other fields.get_employee- full employee card by id.get_employee_by_email- full employee card by email.
Goals and OKR
Section titled “Goals and OKR”search_goals- search goals with filters.get_goal- get a goal with its key results.create_goal- create a goal.update_goal- update a goal.create_key_result- add a key result to a goal.update_key_result- update a key result.checkin_goal- post a check-in with current progress.
Goals and key results cannot be deleted through the assistant.
Time off
Section titled “Time off”search_time_off_calendar- who is absent in a period across the organization, a department, or a team. Answers questions like “who is off next week” in one step.get_employee_time_off_calendar- one employee’s time off for a year.search_employee_time_off- an employee’s time-off history: accruals, usage, and requests.get_employee_time_off_balance- an employee’s balance for a category over a year.list_employee_time_off_categories- an employee’s time-off categories with available balance.
Private time-off categories are never shown in the calendar.
Surveys
Section titled “Surveys”search_surveys- surveys you can see, with status, dates, and completion.get_survey_results- results by question: answer distributions, eNPS with breakdowns by department and manager, and a preview of open-ended answers.search_survey_answers- full open-ended answers to one question, page by page.search_survey_participants- everyone invited to a survey with their completion status. Answers questions like “who has not completed the survey yet”. On anonymous surveys completion status is not shown.
Performance reviews
Section titled “Performance reviews”search_performance_reviews- review cycles with status, dates, and completion.get_performance_review_stats- how a cycle is going: how many reviews are invited, in progress, and completed, overall and by reviewer type.get_performance_review_results- results of one reviewer type (for example, peers) across the whole cycle.get_performance_review_participant_report- scores and comments for one participant by reviewer type, for example to prepare for a one-on-one.
Anonymity
Section titled “Anonymity”Anonymous surveys and anonymous reviewer types stay anonymous through the assistant:
- While fewer people have answered than the anonymity threshold set for the survey or reviewer type, results and answers are not shown at all. The assistant tells you that there are not enough answers yet instead of drawing conclusions.
- Answers to anonymous surveys are not linked to people, so breakdowns by department or manager are empty.
- For anonymous reviewer types the assistant sees only the combined score, not individual ratings.
- Surveys and performance reviews are read-only: the assistant cannot create, edit, or take them.
- With write access the assistant can create and update goals and key results. Review what the assistant proposes before confirming such an action.
- To make sure the assistant never changes data, uncheck write access when you connect.