@checked/mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| CHECKED_TOKEN | No | An alternative credential to CHECKED_API_KEY. Can be used instead of a workspace API key or a saved login. | |
| CHECKED_API_KEY | No | A workspace API key (Settings, API keys in the app). Reaches that workspace only. Can be used instead of CHECKED_TOKEN or a saved login. | |
| CHECKED_API_URL | No | Only if you use a server other than https://getchecked.twelveai.app. | https://getchecked.twelveai.app |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| whoamiA | Read-only. Shows which Check’d account or workspace API key this server uses, its workspaces and plans. |
| list_projectsB | Read-only. Lists Check’d projects (API projects test endpoints; website projects are scanned in a browser) with their last run. |
| list_environmentsA | Read-only. Lists a project’s environments (base URL, auth type). Secrets are masked. |
| list_runsC | Read-only. Recent runs of a project, newest first. |
| get_runA | Read-only. A compact result for one run: status and counts, release gates, the AI verdict and the top problems with ids for get_finding. |
| wait_for_runA | Read-only. Waits (at most 600 s) for a queued or running run to finish, then returns its result. Does not start anything. |
| get_run_resultsB | Read-only. Individual API test results of a run; filter to failures with status "failed,errored". |
| get_run_reportA | Read-only. The AI QA report for a run: verdict (ship, ship with care, hold), findings with impact and fixes, next steps, what was not tested. |
| list_findingsA | Read-only. Problems from a run (default the latest finished one): failed API tests and website scan findings, worst first. |
| get_findingA | Read-only. One finding with its full evidence: the exact request and response, failed checks, stack trace or console errors when captured, the diagnosis and how to fix it. |
| list_issuesA | Read-only. Open issues for a project from the issue inbox: errors real users hit in production and findings from tests and scans, highest impact first, with the one-line AI summary when one exists. On servers without the inbox, the latest run’s findings. |
| get_issueA | Read-only. One issue: occurrences, users, releases, the privacy-safe evidence (error, stack frames in the app’s code, breadcrumbs, request), its fix pull request, and the cached AI explanation (what is broken, likely cause, where to look, suggested fix). Never starts a new explanation. Also accepts finding ids (r123, f456). |
| get_usageB | Read-only. How much of this session’s and this month’s allowance the workspace has used, and whether new work is blocked. |
| preview_importA | Read-only (nothing is saved). Parses an OpenAPI, Swagger, Postman, Insomnia, HAR, GraphQL or cURL document, given as text or a URL, and says what an import into the project would add. |
| import_specA | Saves endpoints into an API project from an OpenAPI, Postman or other document (merged by method and path; replace swaps the whole list). Call preview_import first and confirm with the user. May start an AI reading of the API, which uses allowance. |
| start_runA | Starts a run on a project: API tests for API projects, a website scan (or AI exploration) for website projects. Uses the workspace’s usage allowance (session and month). Optionally waits for the result (at most 600 s); otherwise use wait_for_run or get_run. |
| rerun_failedA | Re-runs only the failed and errored tests of a finished API run, in place, to confirm a fix. Uses the workspace’s usage allowance (session and month). |
| start_scanA | Scans a website for broken links, console and network errors, accessibility and speed. Uses the workspace’s website project for that host, creating one if needed. Uses the workspace’s usage allowance (session and month). |
| cancel_runA | Cancels a queued or running run. Stops further usage by that run. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| fix_failures | Look at the latest run of a project, explain the failures and fix them in this codebase. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 19 tools
Most tools have clearly distinct resources/actions (projects, environments, runs, findings, issues, imports, scans, usage). A few could be confused—start_run vs start_scan on website projects, and list_issues/list_findings/get_issue/get_finding around problem retrieval—but the detailed descriptions explain the differences.
All tool names are snake_case and mostly follow a verb_noun pattern (list_projects, get_run, start_run, cancel_run). Minor deviations such as whoami, wait_for_run, and rerun_failed keep it from perfect consistency.
19 tools is in the heavy 16-25 range for a single integration. While most map to distinct entities, several granular run/finding retrieval tools (get_run_results, get_run_report, list_findings, get_finding) could be consolidated.
Core operational workflows are covered: inspect projects/environments/runs/findings/issues, import specs, start/rerun/cancel runs, scan websites, and check usage. However there are notable gaps in setup/admin—no create/update/delete for projects or environments, and issue/finding tools are read-only—so onboarding a new API project or editing configuration is a dead end.