MiniMax-Code-Plugins/.github/workflows/tool-map-windows.yml at main · MiniMax-AI/MiniMax-Code-Plugins · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
name: tool-map (windows-latest)

on:
  pull_request:
    paths:
      - 'plugins/antianqi/tool-map/**'
      - '.github/workflows/tool-map-windows.yml'
      - 'test/tool-map.test.mjs'
  push:
    branches: [main]
    paths:
      - 'plugins/antianqi/tool-map/**'
      - '.github/workflows/tool-map-windows.yml'
      - 'test/tool-map.test.mjs'
  # Manual dispatch: lets a maintainer / the PR author trigger the
  # same windows-latest job outside a PR. Used to capture a
  # github-hosted green check on the fork (the fork-to-upstream PR
  # itself cannot trigger Actions without explicit maintainer
  # approval, and first-time-contributor protection is on).
  workflow_dispatch:

# PR #5 round-6 review (hetaoBackend, 2026-09-01T01:24:53Z) on commit
# 6bb6a4b: "POSIX tests pass 29/29 ... The remaining blocker is
# platform evidence: the Windows/.cmd/.bat tests return early on
# non-Windows, and this head has no GitHub Actions run, so the new
# windows-latest workflow has not actually validated the
# shell/PATHEXT path. Please provide a real Windows run before merge."
#
# This workflow exercises the existing test/tool-map.test.mjs on
# windows-latest. The two test cases gated on win32 are:
#
#   - "Windows: probeVersion handles the PATHEXT-expanded .CMD
#     path (R4-4 real Windows evidence)" -- creates a fake
#     `node.cmd` in a temp dir, sets PATH, asserts probeVersion
#     resolves the .cmd shim and captures `node core` via the
#     PATHEXT lookup. This is the only line of code that decides
#     whether a .cmd shim routes through cmd.exe (CVE-2024-27980) or
#     spawns as a normal executable.
#   - "shouldUseShell agrees with shellForFile for every whitelisted
#     probe that is installed" -- runs scan.mjs's `shouldUseShell`
#     against the installed tools on the runner and asserts the
#     decision is consistent with the resolved file extension.
#
# These two tests were SKIPPED on every previous CI run (POSIX
# runner); this workflow is what makes the round-6 "real Windows
# run" requirement reproducible in CI. The local-runnable mirror
# `plugins/antianqi/tool-map/scripts/test-windows-workflow-local.ps1`
# gives the same evidence without requiring Actions approval from
# the maintainer.
#
# `[code]smith` is SKIPPED on this repository, so this windows-latest
# job is the CI evidence for the round-6 review.

permissions:
  contents: read

jobs:
  tool-map-windows:
    name: tool-map on windows-latest (.cmd/.bat / PATHEXT / shell)
    runs-on: windows-latest
    timeout-minutes: 10
    defaults:
      run:
        shell: pwsh
    steps:
      - name: Checkout
        uses: actions/checkout@v4

      # Use the system Node so the .cmd / .bat PATHEXT lookup uses the
      # same Node version the reviewer tested against. The runner
      # images ship with Node 20.x as of 2026-09-01.
      - name: Set up Node (system)
        run: |
          node --version
          npm --version

      # Round-6 reviewer finding: ".cmd/.bat tests return early on
      # non-Windows." On windows-latest the if (process.platform
      # !== 'win32') return guards in the test bodies will NOT trip,
      # and the R4-4 .cmd / .bat evidence will actually exercise.
      - name: Run tool-map Windows test suite
        run: |
          cd '${{ github.workspace }}'
          node --test test/tool-map.test.mjs