Repository navigation
Expand file tree
/
Copy pathpnpm-workspace.yaml
More file actions
214 lines (206 loc) · 10.6 KB
/
Copy pathpnpm-workspace.yaml
File metadata and controls
214 lines (206 loc) · 10.6 KB
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
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
packages:
- 'packages/*'
- 'apps/*'
# Supply-chain hardening — see SECURITY.md for rationale.
# Cooldown unit is minutes; 10080 = 7 days.
minimumReleaseAge: 10080
# Excluded: user-authored packages (no security benefit from cooldown) +
# high-trust org scopes that release lockstep (would block install for a
# week on every dep bump). The trailing entries are temporary — see TODOs.
minimumReleaseAgeExclude:
- brepjs
- brepkit-wasm
- occt-wasm
- '@types/*'
- '@typescript-eslint/*'
- typescript-eslint
# @typescript/native-preview (the tsgo native compiler, incl. its
# per-platform binary packages) publishes dated dev builds daily, so a fixed
# cooldown would perpetually block the newest one.
- '@typescript/native-preview*'
- '@posthog/*'
- posthog-js
- '@vercel/*'
- '@protobufjs/*'
- protobufjs
# Security fixes younger than the window, each pinned to the exact reviewed
# version — never a bare package name, which would drop the cooldown for
# every future release of it. One package needing two versions exempted takes
# a single `||` union entry, never a second line (see SECURITY.md).
#
# fast-uri 3.1.9 (2026-10-19) fixes GHSA-w74x-v6f9-j3xg, GHSA-xhq7-xmqj-769h,
# GHSA-wjr9-j795-jxcr and GHSA-333x-w4r6-6crp. Reached only through ajv
# (PWA build, @vercel/node). TODO: remove after 2026-10-27.
- fast-uri@3.1.9
# serialize-javascript 7.1.3 (2026-10-22) fixes GHSA-m5mm-p8h7-m29v. Reached
# only through workbox-build (PWA build). TODO: remove after 2026-10-30.
- serialize-javascript@7.1.3
# minimatch 11.0.3 and glob 14.0.3 (2026-10-28) fix GHSA-g3hv-75wg-m249 and
# GHSA-3g3m-2fjr-6pw4, neither backported to an older major. path-scurry 3.0.0
# (2026-10-27) and @isaacs/nonrecursive-glob-match 1.0.3 (2026-10-28) arrive
# as their dependencies. TODO: remove after 2026-11-04.
- minimatch@11.0.3
- glob@14.0.3
- path-scurry@3.0.0
- '@isaacs/nonrecursive-glob-match@1.0.3'
# Blocks Shai-Hulud-class worms that spread via npm lifecycle scripts.
# Packages that need to build are listed in allowBuilds.
ignoreScripts: true
allowBuilds:
core-js: true
esbuild: true
protobufjs: true
unrs-resolver: true
peerDependencyRules:
allowedVersions:
vite-plugin-pwa>vite: '^8.0.0'
# A security override is a FLOOR, and its selector has to say which major line
# it floors. A bare `pkg: '^X'` matches every major, so a consumer that asks for
# a newer major gets silently dragged DOWN to X; an unbounded `pkg@<X` matches
# every older major, so an old consumer gets dragged UP across a breaking
# boundary. Both read as ordinary pins and neither fails loudly. Range-scope per
# major instead, one entry per affected line, and let OSV flag a major that has
# no rule yet rather than guessing at one. `parent>child` entries are exempt --
# naming the parent already bounds the blast radius.
overrides:
'@vercel/node>undici': '^7.30.1'
# GHSA-xjj6-f33q-929c and GHSA-4qv8-2p9j-2685 are fixed at 6.29.1, 7.30.1
# and 8.12.2 on their lines; GHSA-f72v-6w45-q472 at 7.30.1 and 8.12.2.
'undici@>=6.0.0 <6.29.1': '^6.29.1'
'undici@>=7.0.0 <7.30.1': '^7.30.1'
'undici@>=8.0.0 <8.12.2': '^8.12.2'
# GHSA-mrhx-c7xc-m3w5 is fixed at 4.6.6. @vercel/cli-config (under
# @vercel/blob) pins 4.1.11 exactly.
'zod@>=4.0.0 <4.6.6': '^4.6.6'
# GHSA-9wv6-86v2-598j — ReDoS, fixed at 6.3.0 for the 4.x-6.x range with no
# backport. Scoped to 6.x, the only line in the tree (@vercel/node pins 6.1.0).
# The 7.x and 8.x lines have their own advisories and their own fixes; a bare
# rule here would have pinned either of them back to 6.3.0.
'path-to-regexp@>=6.0.0 <6.3.0': '6.3.0'
'@rollup/plugin-replace>magic-string': '0.30.21'
'@surma/rollup-plugin-off-main-thread>magic-string': '0.30.21'
'workbox-build>@rollup/plugin-node-resolve': '^16.0.0'
'@apideck/better-ajv-errors>ajv': '^8.17.1'
# ajv pins ^3.0.1; the 4.x line has separate advisories fixed at 4.1.2.
'fast-uri@>=3.0.0 <3.1.9': '^3.1.9'
# sharp <0.35.0 inherits libvips CVE-2026-33327/33328/35590/35591
# (GHSA-f88m-g3jw-g9cj). Only reachable transitively via
# manifold-3d>@gltf-transform/functions>ndarray-pixels (build-time glTF
# texture processing, never bundled into the browser). ndarray-pixels pins
# sharp ^0.34.0; there is no patched 0.34.x, so force the 0.35 line.
sharp: '^0.35.0'
# Floor raised 7.0.3 -> 7.0.5: GHSA-qj8w-gfj5-8c6v affects [5.0.0, 7.0.5), so
# the old floor was itself a vulnerable version. It never resolved there --
# @rollup/plugin-terser asks for ^7.0.3 and gets 7.0.7 -- so OSV stayed green
# and the stale floor never surfaced.
# 7.1.2 fixes GHSA-gfhx-hw2g-v5hg; 7.1.3 fixes GHSA-m5mm-p8h7-m29v.
'serialize-javascript@>=7.0.0 <7.1.3': '^7.1.3'
'eslint-plugin-react-hooks>eslint': '^10.0.3'
'eslint-plugin-jsx-a11y>eslint': '^10.0.3'
# GHSA-g3hv-75wg-m249 covers every minimatch below 11.0.1 with no backport, so
# each consumer is moved to 11.x by name. 11.0 removed makeRe(),
# Minimatch#regexp and the AST class; every consumer below calls only
# minimatch(), minimatch.match() or new Minimatch().match(), which 11 keeps. A
# consumer that needs the removed API cannot join this list. eslint's file walk
# depends on Minimatch#match(path, true), whose `**` partial matching changed
# in 11.0, so a bump here needs lint parity proven, not just a green lint.
#
# jsx-a11y and @ts-morph/common call minimatch's *default* export, which no
# CJS build since 9.0 defines, so both are patched (patches/) to import the
# named export.
'eslint>minimatch': '^11.0.3'
'@eslint/config-array>minimatch': '^11.0.3'
'@typescript-eslint/typescript-estree>minimatch': '^11.0.3'
'eslint-plugin-jsx-a11y>minimatch': '^11.0.3'
'@ts-morph/common>minimatch': '^11.0.3'
'filelist>minimatch': '^11.0.3'
'@vercel/static-config>ajv': '^8.18.0'
# GHSA-3g3m-2fjr-6pw4 covers glob 9.0.0 through 14.0.1. Both consumers call
# glob() or globSync() with options 14 keeps.
'workbox-build>glob': '^14.0.3'
'@vercel/nft>glob': '^14.0.3'
'flatted@>=3.0.0 <3.4.2': '^3.4.2'
# posthog-js (^3.3.2) and @cesium/engine (^3.3.0) both sit on 3.x.
# 3.4.16 fixes GHSA-p98j-92pf-mc4p.
'dompurify@>=3.0.0 <3.4.16': '^3.4.16'
'@babel/core@>=7.0.0 <7.29.6': '^7.29.6'
# GHSA-5p4m-2wfm-xmqj (quadratic CPU in !!omap resolution) is fixed in 4.3.1
# on the 4.x line and 3.15.1 on 3.x, with no single version covering both.
# Range-scoped per line so a 3.x consumer is bumped within its own major —
# the previous bare `js-yaml@<4.2.0` rule matched 3.x too and would have
# forced it across a major boundary. Only 4.x is in the tree today; the 3.x
# rule is inert until something pulls it in.
'js-yaml@>=4.0.0 <4.3.1': '^4.3.1'
'js-yaml@>=3.0.0 <3.15.1': '^3.15.1'
# GHSA-2v37-7h3g-55p8 / CVE-2026-67213 — infinite loop when a custom generator
# gets size 0.
# Arrives via postcss. Lower-bounded for the same reason as js-yaml above: the
# advisory's range opens at 0, so an unbounded `<3.3.18` would drag a 1.x/2.x
# consumer across two majors. There is no backport below 3.3.18, so an older
# major cannot be patched in place anyway — OSV would flag it and it gets its
# own rule then. The 4.x–5.x affected range is not in the tree.
'nanoid@>=3.0.0 <3.3.18': '^3.3.18'
# 1.2.2 fixes GHSA-68fv-2mgg-jv7q.
'source-map-js@>=1.0.0 <1.2.2': '^1.2.2'
# 4.7.10 fixes GHSA-8r5x-fm3f-whwj, GHSA-p8wg-vrv2-v86f and
# GHSA-xw65-4hp5-5hc7. eslint-plugin-boundaries pins 4.7.9 exactly.
'handlebars@>=4.0.0 <4.7.10': '^4.7.10'
# The bare `protobufjs: ^7.6.4` this replaces was not latent -- it was firing.
# @cesium/engine asks for ^8.6.5 and was being pinned back to the 7.x line,
# two majors of API behind what it declares, for no security gain: 8.x resolves
# above its own last advisory (8.6.6) and is clean. Splitting per line lets
# cesium have the major it asked for and keeps a floor on each.
#
# The 7.x floor also moves 7.6.4 -> 7.6.5: GHSA-j3f2-48v5-ccww affects
# [7.5.0, 7.6.5), so the old floor was a vulnerable version that only stayed
# green because `^` resolved past it.
'protobufjs@>=7.0.0 <7.6.5': '^7.6.5'
'protobufjs@>=8.0.0 <8.6.6': '^8.6.6'
'eslint-plugin-i18next>lodash': '^4.18.1'
'vitest-axe>lodash-es': '^4.18.1'
'picomatch@>=4.0.0 <4.0.4': '^4.0.4'
# brace-expansion arrives only via minimatch@11 now (all consumers pinned to
# 11.x above), which uses the 5.x line. Floor at 5.0.12, the first release
# patched against GHSA-rgw5-rvv9-x895, GHSA-6j4f-fj2g-mc7p, GHSA-qhr7-859c-m2p7
# and GHSA-q2hr-2g5m-vwhr. The 1.x line is gone, so its old
# `brace-expansion@<1.1.17` pin is no longer needed (#2974).
'brace-expansion@>=5.0.0 <5.0.12': '^5.0.12'
# @mapbox/node-pre-gyp asks for ^7.4.0. Lower-bounded because tar 6.x is very
# common transitively and an unbounded `<7.5.21` would have forced any that
# appeared onto 7.x, which is an ESM/API break.
'tar@>=7.0.0 <7.5.21': '^7.5.21'
# GHSA-8g5g-mxqh-grgj is fixed at 6.4.5. Parent-scoped because a
# `vite@>=6.0.0 <6.4.5` selector also matches every plugin's vite peer range
# and rewrites it to ^6.4.5, under an app that runs vite 8.
'@playwright/experimental-ct-core>vite': '^6.4.5'
'@vercel/python-analysis>smol-toml': '^1.6.1'
esbuild: '0.28.1'
# GHSA-vfj7-8cjw-p6xm (braces <= 3.0.3: unbounded recursion on nested
# patterns) has no patched release, so braces is removed instead, by cutting
# the micromatch@4 and fast-glob edges that pulled it in. The consumers are
# patched onto picomatch@2 and tinyglobby (packageExtensions below). A new
# micromatch consumer reintroduces braces and OSV goes red again.
'@boundaries/elements>micromatch': '-'
'eslint-plugin-boundaries>micromatch': '-'
'@ts-morph/common>fast-glob': '-'
# GHSA-gcqx-w48m-w2h4 (deepmerge <= 4.3.1: prototype poisoning) has no
# patched release. Its only consumer deep-copies a constant of string arrays
# and booleans, so the patch swaps that one call for structuredClone.
'@rollup/plugin-node-resolve>deepmerge': '-'
# @boundaries/elements calls only isMatch, capture and makeRe, which
# micromatch@4.0.8 forwards to picomatch@2 unchanged, so its matching semantics
# are identical.
packageExtensions:
'@boundaries/elements@3.1.1':
dependencies:
picomatch: '^2.3.2'
'@ts-morph/common@0.11.1':
dependencies:
tinyglobby: '^0.2.17'
patchedDependencies:
'@boundaries/elements@3.1.1': patches/@boundaries__elements@3.1.1.patch
'@rollup/plugin-node-resolve@16.0.3': patches/@rollup__plugin-node-resolve@16.0.3.patch
'@ts-morph/common@0.11.1': patches/@ts-morph__common@0.11.1.patch
brepjs@20.5.0: patches/brepjs@20.5.0.patch
eslint-plugin-jsx-a11y@6.10.2: patches/eslint-plugin-jsx-a11y@6.10.2.patch
three-stdlib: patches/three-stdlib.patch