Cube MCP (edgescaleai)
Capped at 70: no passing tool call yet; the tests add up to 90. Why scores are capped
Test points add up to 90; the score is capped at 70 until a tool call passes. Not a quality or security rating. Methodology
Test points before the cap (90/100)
- Install20/20
- Handshake30/30
- Maintenance25/25
- Docs5/15
- License10/10
Tested npm @edgescaleai/cube-mcp 0.3.66 (MIT; repo link in package metadata not found on GitHub (github.com returns HTTP 404, rechecked 2026-11-18 UTC; we link the npm package page instead)). Installed pinned with scripts off (npm --ignore-scripts) and launched: initialize and tools/list over stdio succeeded with no credentials and listed 45 tools (no tool annotations). Real calls: 0 of 2 real function calls passed. Calls need a CUBE_API_KEY. Core tools not tested: 70 cap.
Tool-call test (2026-11-18 UTC, direct MCP stdio client, no LLM): fail. 0 of 2 real function calls passed. Calls need a CUBE_API_KEY. Evidence: local test run, Nov 18, 2026 UTC.
Transport (tested 2026-11-18 UTC): stdio
Token cost (estimate): approx. 8252 tokens, cl100k estimate (45 tools, measured 2026-11-18 UTC, version 0.3.66). This is an estimate from our testing of the tool names, descriptions and input schemas, counted with cl100k_base.
Permissions: 19 read, 18 write (state-changing), 8 unclassified · Auth: API key · Rate limit: unknown · Pricing unit: unknown
Auth: API key. Rate limit: unknown. Pricing unit: unknown.
acr_get_token— readagent_login_browser— unclassifiedagent_logout— write (state-changing)agent_status— readapp_list— readapp_proxy— unclassifiedapp_proxy_status— readapp_proxy_stop— write (state-changing)build_and_publish_to_apollo— write (state-changing)build_image_prepare— unclassifiedbuild_image_start— write (state-changing)build_image_status— readcancel_active_plan— write (state-changing)compare_product_versions— unclassifiedcreate_secret— write (state-changing)cube_list— readcube_status— readenforce_entity_config— write (state-changing)entity_health— unclassifiedexpose— write (state-changing)get_product_releases— readgroundcover_k8s_events— unclassifiedgroundcover_list_clusters— readgroundcover_list_metrics— readgroundcover_list_workloads— readgroundcover_query_metrics— unclassifiedgroundcover_search_logs— readgroundcover_search_traces— readinstall_entity— write (state-changing)install_module— write (state-changing)kubectl_exec— write (state-changing)list_environments— readlist_exposed— readlist_modules— readlist_products— readlist_release_channels— readlist_services— readplan_details— unclassifiedsend_feedback— write (state-changing)unexpose— write (state-changing)uninstall_entity— write (state-changing)uninstall_module— write (state-changing)update_entity_config— write (state-changing)update_module_variables— write (state-changing)update_secret— write (state-changing)
listing-flags calls_need_key or auth_kind
Score 70/100 · Tools 45 · Token cost approx. 8,252 tokens · Compare · Last tested 2026-11-18 UTC · Version tested 0.3.66 · Package @edgescaleai/cube-mcp
Summary: Cube MCP (edgescaleai) completed an MCP handshake without credentials and exposed 45 tools.
Evidence summary
- Handshake
- pass
- Tools listed
- 45
- Context cost
- approx. 8252 tokens, cl100k estimate (45 tools, measured 2026-11-18 UTC, version 0.3.66; an estimate) ranking · add up your servers' context cost
- API key
- Connects without a key; tool calls need one
- Tested
- 2026-11-18 UTC
Test results
| Test | Result | Points | Evidence |
|---|---|---|---|
| install | pass | 20 | npm install @edgescaleai/cube-mcp --ignore-scripts (Node v24.22.1, Linux) succeeded in 0s. |
| handshake | pass | 30 | MCP initialize + tools/list over stdio succeeded with no credentials; 45 tools listed in 12.9s. |
| maintenance | pass | 25 | Latest release published 1 days before test. |
| docs | partial | 5 | install instructions present. Check of the package's own README (README shipped in npm package). |
| license | pass | 10 | MIT (OSI-approved). |
Local-model readiness
Token cost (estimate): approx. 8252 tokens, cl100k estimate for 45 tools · about 101% of an 8k context. This is an estimate from our testing (counted with cl100k_base).
Small-model friendly (schema heuristic, not a model result): unknown.
Not tested
Small-model task runs: Not yet tested.
What we tested and what we didn't
Tested on 2026-11-18 UTC
- Install from the registry
- We connected with no credentials and asked for its tool list: it listed 45 tools (Method details: MCP
initializethentools/list) - Maintenance signals (last push, archived/deprecated status)
- Docs (README length, install and config example)
- License (OSI-approved or not)
- Token cost (estimate)
Not tested
- Security: not audited (no code review, no permission or data-flow analysis)
- Real tasks (no functional task runs yet)
- Behaviour with real credentials or a real account
- Performance, rate limits or load
- Correctness of results beyond the listed tasks
Methodology: how we test. A high score means it installs and responds, not that giving it access to your data carries no risk.
Known security issues
Not yet checked.
Monthly roundup: MCP security issues, October 2026.
Tool list changes
Baseline recorded from the 2026-11-18 UTC test run; changes appear after the next test.
Note: tool list can depend on the connected database, so changes are annotated, not flagged automatically.
Snapshot hash (fingerprint of the tool list)
87d610b7669e: SHA-256 of the canonical tool names, descriptions and input schemas. If it changes, the tool list changed.
Raw test record (JSON)
{
"name": "@edgescaleai/cube-mcp",
"version": "0.3.66",
"registry": "npm",
"status": "working",
"tested": "2026-11-18",
"tools_count": 45,
"deprecated": false,
"archived": null,
"tests": {
"install": [
"pass",
20,
"`npm install @edgescaleai/cube-mcp --ignore-scripts` (Node v24.22.1, Linux) succeeded in 0s."
],
"handshake": [
"pass",
30,
"MCP `initialize` + `tools/list` over stdio succeeded with no credentials; 45 tools listed in 12.9s."
],
"maintenance": [
"pass",
25,
"Latest release published 1 days before test."
],
"docs": [
"partial",
5,
"install instructions present. Check of the package's own README (README shipped in npm package)."
],
"license": [
"pass",
10,
"MIT (OSI-approved)."
]
},
"testedAt": "2026-11-18T04:18:16.857Z"
}Install
npx -y @edgescaleai/cube-mcp
Pin the version we tested
npx -y --ignore-scripts @edgescaleai/cube-mcp@0.3.66
tested 2026-11-18 UTC. Newer versions may differ and were not tested by us.
{
"mcpServers": {
"cube-mcp-edgescaleai": {
"command": "npx",
"args": [
"-y",
"@edgescaleai/cube-mcp@0.3.66"
]
}
}
}Tools listed
Showing 40 of 45 names: list_environments, install_entity, entity_health, plan_details, cancel_active_plan, acr_get_token, build_image_prepare, build_image_start, build_image_status, uninstall_entity, update_entity_config, enforce_entity_config, list_modules, install_module, uninstall_module, update_module_variables, expose, unexpose, list_exposed, create_secret, update_secret, list_products, compare_product_versions, list_release_channels, get_product_releases, cube_status, cube_list, kubectl_exec, app_list, list_services, groundcover_list_clusters, groundcover_list_workloads, groundcover_query_metrics, groundcover_list_metrics, groundcover_search_logs, groundcover_search_traces, groundcover_k8s_events, send_feedback, build_and_publish_to_apollo, app_proxy.Show all 45 names
list_environments, install_entity, entity_health, plan_details, cancel_active_plan, acr_get_token, build_image_prepare, build_image_start, build_image_status, uninstall_entity, update_entity_config, enforce_entity_config, list_modules, install_module, uninstall_module, update_module_variables, expose, unexpose, list_exposed, create_secret, update_secret, list_products, compare_product_versions, list_release_channels, get_product_releases, cube_status, cube_list, kubectl_exec, app_list, list_services, groundcover_list_clusters, groundcover_list_workloads, groundcover_query_metrics, groundcover_list_metrics, groundcover_search_logs, groundcover_search_traces, groundcover_k8s_events, send_feedback, build_and_publish_to_apollo, app_proxy, app_proxy_stop, app_proxy_status, agent_login_browser, agent_logout, agent_status
Quick config
{
"mcpServers": {
"cube-mcp-edgescaleai": {
"command": "npx",
"args": [
"-y",
"@edgescaleai/cube-mcp"
]
}
}
}Add to config Opens the config generator with this server selected.
Setup
- Install. Run
npm install @edgescaleai/cube-mcp --ignore-scripts(Node v24.22.1). This is the exact command we ran on 2026-11-18 UTC. - Configure. Add the mcpServers entry below to your MCP client config.
- Verify. Restart the client and check that tools/list returns tools (45 in our test).
{
"mcpServers": {
"cube-mcp-edgescaleai": {
"command": "npx",
"args": [
"-y",
"@edgescaleai/cube-mcp"
]
}
}
}Test history
- 2026-11-18 UTC: New
FAQ
Is Cube MCP (edgescaleai) deprecated?
No. Cube MCP (edgescaleai) was not deprecated or archived when we last tested it on 2026-11-18 UTC.
What replaces Cube MCP (edgescaleai)?
Nothing; Cube MCP (edgescaleai) is not deprecated as of 2026-11-18 UTC.
Does Cube MCP (edgescaleai) need an API key?
It connects and lists 45 tools without a key, but its tool calls need one.
Embed the badge
Every tested server gets a free embeddable badge. It shows the current score and test date and expires after 90 days. Testing is free; payment never changes a score.