AI agents call gitlab_action to retrieve information from UnClick without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
| Parameter | Type | Required | Description |
|---|---|---|---|
page | number | — | Page number. |
query | string | — | Search query string (for search_projects). |
state | string | — | Filter by state: opened, closed, merged. |
action | string | Yes | Action: search_projects, get_project, list_issues, list_mrs, get_user. |
labels | string | — | Comma-separated label names to filter by. |
base_url | string | — | GitLab base URL (default: https://gitlab.com/api/v4). Set for self-hosted instances. |
per_page | number | — | Results per page. |
username | string | — | GitLab username (for get_user). |
project_id | string | — | Project ID or URL-encoded namespace/project path. |
access_token | string | — | GitLab personal access token (PAT). |
Parameters from the server's own tool schema.
gitlab_action is narrowly scoped to read operations against GitLab's REST API. It retrieves existing data (projects, issues, merge requests, user information) without the ability to create, modify, or delete resources. The only modest risk is potential information disclosure if an agent queries sensitive project details it shouldn't access, but this depends on API token permissions rather than the tool itself.
From the tool's definition The tool description explicitly states it performs read-only operations: 'search projects, get project details, list issues and merge requests, and look up users.' These are all query/retrieval actions with no mention of modifications, deletions, or side…
Risk signalsAccepts freeform code/query input (query) · Accepts URL/endpoint input (base_url) · Handles credentials or secrets (access_token) · High parameter count (10 properties)
Attacks that exploit this kind of access
Interact with the GitLab REST API: search projects, get project details, list issues and merge requests, and look up users. Supports self-hosted GitLab instances. It is categorised as a Read tool in the UnClick MCP Server, which means it retrieves data without modifying state.
gitlab_action accepts 10 parameters: page, query, state, action, labels, base_url, per_page, username, project_id, access_token. Required: action. The full parameter table on this page comes from the server's own tool schema.
Register the UnClick MCP server in PolicyLayer and add a rule for gitlab_action: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches UnClick. Nothing to install.
gitlab_action is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the gitlab_action rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for gitlab_action. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
gitlab_action is provided by the UnClick MCP server (@unclick/mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →