You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
TUI real-terminal smoke test self-skips in CI — regressions ship with zero CI signal #2408
v2 — current (@modelcontextprotocol/inspector@latest)
Which client?
TUI
Inspector version
2.7.0 (git tag) — static code-review finding, not run locally
Node version
N/A — static code review, no live run performed
Operating system (and browser, for the web client)
N/A — static code review
Transport
Not applicable / never connected
MCP server under inspection
N/A — this is a static code-review finding against the 2.7.0 tag's source and CI config (docs/quality-gate.md), not a live reproduction against a running MCP server.
Steps to reproduce
Found via static review of the 2.7.0 tag source (docs/quality-gate.md), not a live run.
docs/quality-gate.md states that smoke:tui self-skips under process.env.CI because Ink needs a real TTY.
It only runs "for real" during npm run local:gate on a developer's own machine.
Confirmed this is the documented, intentional behavior — not a flake.
I have not attempted to push a TUI regression through CI to watch it slip through live; flagging based on the documented gap itself.
Expected behavior
The TUI smoke test runs against a real PTY in CI (e.g. via node-pty or a similar terminal emulator driving scripted key sequences), so a TUI regression is caught before merge/release regardless of whether a contributor happened to run local:gate locally.
Actual behavior
smoke:tui self-skips whenever process.env.CI is set, so CI runs currently give zero signal on TUI regressions; the only real execution path is a developer manually running npm run local:gate.
Suggested fix: run the TUI smoke test in CI against a real PTY instead of relying on local:gate as the only real execution path. By the project's own docs, this is the largest gap between what's tested and what ships.
Logs, errors, or screenshots
No response
Already prototyped a fix?
No response
Before you submit
I searched existing issues and this is not a duplicate.
This is not a security vulnerability report (those go through the private advisory process).
Severity 3 — a CI signal gap, partially covered: smoke:tui does run for real in the mandatory npm run local:gate, so the surface is not unguarded, only unguarded in CI
Urgency 2 — wanted eventually
Bonuses: none
Labels applied at triage: v2, chore (CI/test tooling, not a product defect) — replacing the form's static bug.
Board: #28, Status Incoming (awaiting maintainer review — no milestone yet).
Shipped in #2489 (merged to v2/main as 56ba766). smoke:tui now runs for real in GitHub CI — first runner pass: rendered "MCP Servers" at 434ms and survived 2000ms.
Activity
cliffhall commentedon Sep 18, 2026
Triage: Priority Low (total 5)
smoke:tuidoes run for real in the mandatorynpm run local:gate, so the surface is not unguarded, only unguarded in CILabels applied at triage:
v2,chore(CI/test tooling, not a product defect) — replacing the form's staticbug.Board: #28, Status
Incoming(awaiting maintainer review — no milestone yet).cliffhall commentedon Sep 24, 2026
Shipped in #2489 (merged to
v2/mainas 56ba766).smoke:tuinow runs for real in GitHub CI — first runner pass: rendered "MCP Servers" at 434ms and survived 2000ms.chore(smoke): run smoke:tui for real in GitHub CI (modelcontextprotoc…