jig-ui 0.3.0 → 0.5.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -3,15 +3,22 @@ cite the number when you follow or deliberately break one.
3
3
 
4
4
  ## Before generating or reviewing any UI
5
5
 
6
+ 0. Check that `{{config_file}}` exists. If it does not, this project has no
7
+ token layer: run `init` and stop until it has been run. Do not proceed by
8
+ writing token definitions of your own — a `:root` block you author, however
9
+ clearly you label it, is an invented design system wearing this one's names,
10
+ and it will not be replaced when the real tokens arrive.
6
11
  1. Load `{{rules_path}}/00-anti-patterns.md` and `{{rules_path}}/01-modes.md`.
7
12
  2. Determine the mode from `{{config_file}}`, or infer it using the procedure in
8
13
  `{{rules_path}}/01-modes.md` and state the inference in one line before building.
9
14
  3. Load the relevant section of `{{rules_path}}/03-patterns.md` for the component you are building.
10
15
  4. Load `{{rules_path}}/05-copy.md` whenever you write a label, button, heading, error or empty state.
11
16
  5. Consume tokens by semantic name only. Never write a raw colour or pixel value
12
- at a call site.
13
- 6. Run the self-check at the end of `{{rules_path}}/00-anti-patterns.md` before finishing
14
- (a future `check` command will automate this).
17
+ at a call site, and never resolve a name yourself: if a token you need has no
18
+ value, that is a finding to report, not a number to supply.
19
+ 6. Run `check` for the rules the CLI can detect, then run the self-check at the
20
+ end of `{{rules_path}}/00-anti-patterns.md` for the rules it cannot. Both
21
+ halves, every time — a clean `check` is not a clean review.
15
22
  7. Cite any rule you deliberately break, with the reason, in one line.
16
23
 
17
24
  Load `{{rules_path}}/04-principles.md` only when two rules conflict.
@@ -20,9 +27,15 @@ Load `{{rules_path}}/04-principles.md` only when two rules conflict.
20
27
 
21
28
  {{available_commands}}
22
29
 
23
- Commands marked `available` run via the CLI at `{{scripts_path}}`. Commands
24
- marked `planned` are listed for context only — running one errors out, since
25
- it is not yet registered.
30
+ Commands marked `available` run via the CLI at `{{scripts_path}}` — the version
31
+ this skill was installed with, so the CLI and these rules always agree.
32
+
33
+ `update` is the exception: its job is to move that version forward, so run it
34
+ as `{{update_path}} update`. Run pinned, it refreshes to the version already
35
+ installed and reports success for a no-op.
36
+
37
+ Commands marked `planned` are listed for context only — running one errors out,
38
+ since it is not yet registered.
26
39
 
27
40
  ## Reading command output
28
41
 
@@ -42,12 +55,14 @@ changed since, do not re-run it.
42
55
  Before you finish a task that generated or modified UI, emit this line:
43
56
 
44
57
  ```text
45
- JIG_CHECK: version=<version> mode=<mode> self_check=<ran|skipped>
58
+ JIG_CHECK: version=<version> mode=<mode> mechanical=<pass|fail|skipped>:<n> judgment=<ran|skipped>
46
59
  ```
47
60
 
48
- <!-- `mechanical=<pass|fail>:<n>/<total>` and `judgment=<ran|skipped>` fields
49
- belong here once the `check` command lands and can actually produce
50
- them — do not add them back before that. -->
61
+ `jig check` emits this same line for the half it can do, with
62
+ `judgment=not-run`. Take `mechanical=` from its output rather than counting by
63
+ hand; fill `judgment=` yourself. If `check` could not run at all, that is
64
+ `mechanical=skipped:0` — never `pass`, which would report a clean result for a
65
+ check that never inspected anything.
51
66
 
52
- A skipped self-check must say `skipped` and give the reason. Do not report
53
- `ran` for a check you did not perform.
67
+ A skipped check must say `skipped` and give the reason. Do not report `ran` for
68
+ a check you did not perform.
@@ -15,13 +15,13 @@
15
15
  "status": "available"
16
16
  },
17
17
  "check": {
18
- "description": "Check changed files or a named target against the rule set. Reports violations by rule id.",
19
- "argumentHint": "[target]",
20
- "status": "planned"
18
+ "description": "Check the repo against the mechanical and hybrid rules the CLI can detect. Reports findings by rule id; the judgment rules remain yours to apply.",
19
+ "argumentHint": "[--all] [--ci] [--json]",
20
+ "status": "available"
21
21
  },
22
22
  "explain": {
23
- "description": "Print a rule's full text, its correction, and the version it was introduced in.",
24
- "argumentHint": "<rule-id>",
25
- "status": "planned"
23
+ "description": "Explain a rule, or find the rules you cannot name. Given an id, prints the rule's full text, its correction, the version it arrived in and who checks it — including the P-/M- pattern and mode specs, which no rule index contains. Given a word instead, searches every title and body and lists what matches. `--list` prints every id, or one section's.",
24
+ "argumentHint": "<rule-id | search term> [--list]",
25
+ "status": "available"
26
26
  }
27
27
  }
@@ -99,6 +99,23 @@
99
99
 
100
100
  --color-focus: var(--color-text-strong);
101
101
 
102
+ /* ---- Stroke widths. Options; modes select. ----
103
+ Every rule requiring a visible border or focus ring — E-28, E-29, P-02's
104
+ 3:1 shape floor — used to end at a call site writing `1px` or `2px` by
105
+ hand, which H-47 forbids. Two agents building different components hit
106
+ that independently and reported the same gap: no semantic name existed to
107
+ consume. These are it. */
108
+ --border-width-hairline: 1px; /* control and surface borders */
109
+ --border-width-strong: 2px; /* emphasis, selected states */
110
+
111
+ /* Focus ring geometry. NOT mode-negotiable, for the same reason
112
+ --size-touch-target is not: WCAG 2.4.11 wants a perimeter at least 2px
113
+ thick, and density is never a reason to go under an accessibility floor.
114
+ The offset keeps the ring off the control's own border so both stay
115
+ legible. */
116
+ --focus-ring-width: 2px;
117
+ --focus-ring-offset: 2px;
118
+
102
119
  /* ================= TYPE ================= */
103
120
  --font-text: ui-sans-serif, system-ui, -apple-system, "Segoe UI", sans-serif;
104
121
  --font-display: var(--font-text);
@@ -111,16 +128,33 @@
111
128
  --radius-lg: 32px; /* large surfaces, hero media */
112
129
  --radius-full: 9999px;
113
130
 
114
- /* ================= ELEVATION — two shadows =================
131
+ /* ================= ELEVATION — two shadows, plus none =================
115
132
  Shadow colour is derived from the text colour, not pure black, so it sits in
116
- the palette. Light from above: y-offset positive, blur grows with elevation. */
133
+ the palette. Light from above: y-offset positive, blur grows with elevation.
134
+ `--shadow-none` is the explicit absence a mode selects when its elevation is
135
+ stroke-led; every mode currently does, via `--shadow-surface`. */
117
136
  --shadow-none: none;
118
137
  --shadow-raised: 0 1px 2px rgb(0 0 0 / 6%), 0 2px 4px -2px rgb(0 0 0 / 8%);
119
138
  --shadow-overlay: 0 4px 16px -4px rgb(0 0 0 / 14%), 0 1px 3px rgb(0 0 0 / 8%);
120
139
  }
121
140
 
141
+ /* Dark mode has THREE states, and this file needs two blocks to serve them.
142
+ The media query below covers "the OS says dark", guarded by
143
+ :not([data-theme="light"]) so an explicit light choice still wins. It does
144
+ NOT cover "the OS says light and the user chose dark" — a selector inside a
145
+ media query cannot match when the query is false, so before this second
146
+ block existed, setting data-theme="dark" on a light-mode system produced no
147
+ dark tokens at all. The preview's own dark toggle did nothing there, which is
148
+ why this went unnoticed: dark mode was only ever seen by people whose OS was
149
+ already in it.
150
+
151
+ The two bodies are identical and MUST stay so; check-tokens rule 10 fails the
152
+ build if they diverge. CSS gives no way to share them — a selector list cannot
153
+ straddle a media-query boundary — so the duplication is required and the guard
154
+ is the answer to it. */
122
155
  @media (prefers-color-scheme: dark) {
123
156
  :root:not([data-theme="light"]) {
157
+
124
158
  /* Three solid levels. In dark, DEPTH COMES FROM BACKGROUND, NOT SHADOW —
125
159
  shadows are nearly invisible on a dark surface. Never pure black. */
126
160
  --color-bg-base: hsl(var(--brand-h) 6% 10%);
@@ -150,3 +184,34 @@
150
184
  --shadow-overlay: 0 8px 24px -8px rgb(0 0 0 / 60%);
151
185
  }
152
186
  }
187
+
188
+ :root[data-theme="dark"] {
189
+
190
+ /* Three solid levels. In dark, DEPTH COMES FROM BACKGROUND, NOT SHADOW —
191
+ shadows are nearly invisible on a dark surface. Never pure black. */
192
+ --color-bg-base: hsl(var(--brand-h) 6% 10%);
193
+ --color-bg-raised: hsl(var(--brand-h) 6% 15%);
194
+ --color-bg-overlay: hsl(var(--brand-h) 6% 20%);
195
+
196
+ /* Foregrounds become opacities of white. Higher than the light-mode values,
197
+ because dark interfaces lose detail faster. */
198
+ --color-text-strong: rgb(255 255 255 / 100%);
199
+ --color-text-weak: rgb(255 255 255 / 78%);
200
+ --color-stroke-strong: rgb(255 255 255 / 60%);
201
+ --color-stroke-weak: rgb(255 255 255 / 12%);
202
+ --color-fill: rgb(255 255 255 / 6%);
203
+
204
+ --brand-l: 88%; /* lighten and desaturate the brand for dark surfaces */
205
+ --color-on-brand: hsl(var(--brand-h) 6% 10%);
206
+
207
+ /* System colours lighten and desaturate in dark mode. Only the h/s/l and
208
+ fill-alpha vars need overriding here — the four hsl() declarations
209
+ above pick the new values up automatically. */
210
+ --error-h: 0; --error-s: 60%; --error-l: 72%; --error-fill-a: 8%;
211
+ --warning-h: 42; --warning-s: 70%; --warning-l: 70%; --warning-fill-a: 8%;
212
+ --success-h: 162; --success-s: 45%; --success-l: 66%; --success-fill-a: 8%;
213
+ --info-h: 220; --info-s: 55%; --info-l: 74%; --info-fill-a: 8%;
214
+
215
+ --shadow-raised: none;
216
+ --shadow-overlay: 0 8px 24px -8px rgb(0 0 0 / 60%);
217
+ }
@@ -28,9 +28,27 @@
28
28
  --text-body: 16px; --leading-body: 1.5; /* UI text: labels, controls */
29
29
  --text-prose: 18px; --leading-prose: 1.6; /* LONG-FORM text. Never below 18. */
30
30
  --text-lead: 20px; --leading-lead: 1.5;
31
+
32
+ /* H1 and H2 are FLUID: they interpolate with viewport width instead of
33
+ switching at a breakpoint, which is why this system defines no breakpoint
34
+ token. Each clamp is the line through (360px viewport, min) and (1024px,
35
+ max); outside that range it saturates at the bounds.
36
+
37
+ The `rem` term is not decoration. `clamp(32px, 8vw, 48px)` would ignore a
38
+ reader who has set a larger default font size — page zoom scales `vw`, a
39
+ font-size preference does not — which fails WCAG 1.4.4. Keeping every term
40
+ in `rem` plus `vw` means the whole curve moves with the user's setting.
41
+
42
+ Only headings. Body and prose stay fixed: readability is absolute, set by
43
+ the eye and viewing distance, while heading size is relative — it exists to
44
+ contrast with body, and that contrast can compress on a narrow screen. A
45
+ fluid `--text-prose` would also breach the 18px floor (B-75, T20).
46
+
47
+ The minimums land on rungs the ladder already has rather than on new
48
+ values. Line heights are unitless, so they follow the size for free. */
31
49
  --text-h3: 24px; --leading-h3: 1.333;
32
- --text-h2: 32px; --leading-h2: 1.25;
33
- --text-h1: 48px; --leading-h1: 1.1;
50
+ --text-h2: clamp(1.5rem, 1.229rem + 1.205vw, 2rem); --leading-h2: 1.25; /* 24 → 32 */
51
+ --text-h1: clamp(2rem, 1.458rem + 2.41vw, 3rem); --leading-h1: 1.1; /* 32 → 48 */
34
52
 
35
53
  /* Two weights only. More adds noise and is hard to apply consistently. */
36
54
  --font-weight-regular: 400;
@@ -59,14 +77,39 @@
59
77
  --size-control: 48px; --size-control-sm: 40px;
60
78
  --size-touch-target: 48px; --size-icon: 20px;
61
79
 
80
+ --border-width-control: var(--border-width-hairline);
62
81
  --radius-control: var(--radius-sm); --radius-surface: var(--radius-md);
63
82
  --shadow-surface: var(--shadow-none);
64
83
 
65
84
  --duration-fast: 150ms; --duration-base: 250ms; --duration-slow: 300ms;
66
85
  --ease-out: cubic-bezier(0.16, 1, 0.3, 1);
67
86
  --ease-in-out: cubic-bezier(0.65, 0, 0.35, 1);
87
+
88
+ /* Ambient motion (P-13). Decorative looping movement, editorial only —
89
+ `product` and `operator` define none of these, deliberately.
90
+
91
+ THE THREE VALUES ARE NOT A LADDER, THEY ARE A CHORD. Ambient motion is
92
+ layered: several small movements running at once. If their periods share a
93
+ factor they re-align, and the whole composition pulses in unison — the one
94
+ thing ambient motion must never do. So the periods are mutually prime in
95
+ tenths of a second (30, 47, 71), which puts the first full re-alignment of
96
+ all three past two hours. Round numbers like 3/4/6s would re-align every
97
+ twelve seconds. That is why these look arbitrary; they are not.
98
+
99
+ Pick a different one per layer. Which is which carries no meaning beyond
100
+ speed — unlike the interaction durations above, there is no "this is the
101
+ one for a state change". */
102
+ --duration-ambient-fast: 3s;
103
+ --duration-ambient-base: 4.7s;
104
+ --duration-ambient-slow: 7.1s;
68
105
  }
69
106
 
70
107
  @media (prefers-reduced-motion: reduce) {
108
+ /* The ambient periods are deliberately NOT collapsed to 1ms here. These drive
109
+ infinite loops, so 1ms would spin them at a thousand cycles a second rather
110
+ than stop them — the opposite of what was asked for. P-13 requires the whole
111
+ ambient declaration to sit inside `prefers-reduced-motion: no-preference`,
112
+ so under a reduce preference the animation never starts and there is nothing
113
+ to shorten. Absence, not speed. */
71
114
  :root { --duration-fast: 1ms; --duration-base: 1ms; --duration-slow: 1ms; }
72
115
  }
@@ -27,8 +27,19 @@
27
27
  Line heights UNITLESS. FLOOR: 1.5 holds even here. */
28
28
  --text-caption: 12px; --leading-caption: 1.5;
29
29
  --text-body: 14px; --leading-body: 1.5;
30
- --text-prose: 16px; --leading-prose: 1.6; /* LONG-FORM text only */
30
+ /* 18px, not 16px, and not negotiable by density. `B-75` names 18px as the
31
+ floor for sustained reading, and the reference agrees — "make long body
32
+ text at least 18px". This token sat at 16px, which made our own rule false
33
+ in the one mode most likely to be read for long stretches by someone who
34
+ has been at it all day. Sustained readability is an accessibility floor
35
+ like `--size-touch-target`, not a density dial. Long prose is rare in an
36
+ operator surface; when it appears, it is legible. */
37
+ --text-prose: 18px; --leading-prose: 1.6; /* LONG-FORM text only */
31
38
  --text-lead: 16px; --leading-lead: 1.5;
39
+ /* No fluid headings here, unlike editorial and product. This mode's largest
40
+ heading is 22px, which already fits ~28 characters per line on a 360px
41
+ screen — fluid sizing would have nothing to do. The mechanism is absent
42
+ because the problem is, not by oversight. */
32
43
  --text-h3: 16px; --leading-h3: 1.5;
33
44
  --text-h2: 18px; --leading-h2: 1.4;
34
45
  --text-h1: 22px; --leading-h1: 1.333;
@@ -59,6 +70,7 @@
59
70
  --size-row: 36px; --size-row-compact: 32px;
60
71
  --size-touch-target: 48px; --size-icon: 16px;
61
72
 
73
+ --border-width-control: var(--border-width-hairline);
62
74
  --radius-control: var(--radius-sm); --radius-surface: var(--radius-sm);
63
75
  --shadow-surface: var(--shadow-none);
64
76
 
@@ -66,6 +78,10 @@
66
78
  --ease-out: cubic-bezier(0.2, 0, 0, 1);
67
79
  --ease-in-out: cubic-bezier(0.4, 0, 0.2, 1);
68
80
 
81
+ /* No ambient duration tokens here, deliberately. Ambient motion is
82
+ editorial only (P-13) — a surface someone works in all day must not
83
+ have anything moving on it that they did not cause. */
84
+
69
85
  --font-numeric: var(--font-numeric-features);
70
86
  }
71
87
 
@@ -27,9 +27,27 @@
27
27
  --text-body: 16px; --leading-body: 1.5;
28
28
  --text-prose: 18px; --leading-prose: 1.6; /* LONG-FORM text only */
29
29
  --text-lead: 20px; --leading-lead: 1.5;
30
+
31
+ /* H1 and H2 are FLUID: they interpolate with viewport width instead of
32
+ switching at a breakpoint, which is why this system defines no breakpoint
33
+ token. Each clamp is the line through (360px viewport, min) and (1024px,
34
+ max); outside that range it saturates at the bounds.
35
+
36
+ The `rem` term is not decoration. `clamp(32px, 8vw, 48px)` would ignore a
37
+ reader who has set a larger default font size — page zoom scales `vw`, a
38
+ font-size preference does not — which fails WCAG 1.4.4. Keeping every term
39
+ in `rem` plus `vw` means the whole curve moves with the user's setting.
40
+
41
+ Only headings. Body and prose stay fixed: readability is absolute, set by
42
+ the eye and viewing distance, while heading size is relative — it exists to
43
+ contrast with body, and that contrast can compress on a narrow screen. A
44
+ fluid `--text-prose` would also breach the 18px floor (B-75, T20).
45
+
46
+ The minimums land on rungs the ladder already has rather than on new
47
+ values. Line heights are unitless, so they follow the size for free. */
30
48
  --text-h3: 20px; --leading-h3: 1.4;
31
- --text-h2: 24px; --leading-h2: 1.333;
32
- --text-h1: 32px; --leading-h1: 1.25;
49
+ --text-h2: clamp(1.25rem, 1.114rem + 0.602vw, 1.5rem); --leading-h2: 1.333; /* 20 → 24 */
50
+ --text-h1: clamp(1.5rem, 1.229rem + 1.205vw, 2rem); --leading-h1: 1.25; /* 24 → 32 */
33
51
 
34
52
  --font-weight-regular: 400;
35
53
  --font-weight-bold: 600;
@@ -56,12 +74,17 @@
56
74
  --size-control: 40px; --size-control-sm: 32px; --size-row: 48px;
57
75
  --size-touch-target: 48px; --size-icon: 18px;
58
76
 
77
+ --border-width-control: var(--border-width-hairline);
59
78
  --radius-control: var(--radius-sm); --radius-surface: var(--radius-md);
60
79
  --shadow-surface: var(--shadow-none);
61
80
 
62
81
  --duration-fast: 100ms; --duration-base: 150ms; --duration-slow: 200ms;
63
82
  --ease-out: cubic-bezier(0.16, 1, 0.3, 1);
64
83
  --ease-in-out: cubic-bezier(0.65, 0, 0.35, 1);
84
+
85
+ /* No ambient duration tokens here, deliberately. Ambient motion is
86
+ editorial only (P-13) — a surface someone works in all day must not
87
+ have anything moving on it that they did not cause. */
65
88
  }
66
89
 
67
90
  @media (prefers-reduced-motion: reduce) {