code-factory-4-design 0.3.0__tar.gz

code_factory_4_design-0.3.0/LICENSE-APACHE

@@ -0,0 +1,8 @@
+                                 Apache License
+                           Version 2.0, January 2004
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy at http://www.apache.org/licenses/LICENSE-2.0
+   Distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+   KIND. Full text: http://www.apache.org/licenses/LICENSE-2.0.txt
+   Copyright 2026 WizeMe.APP

code_factory_4_design-0.3.0/LICENSE-MIT

@@ -0,0 +1,10 @@
+MIT License
+
+Copyright (c) 2026 WizeMe.APP
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including the rights to use, copy, modify,
+merge, publish, distribute, sublicense, and/or sell copies, subject to
+including the above notice. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY
+OF ANY KIND. Full text: https://opensource.org/licenses/MIT

code_factory_4_design-0.3.0/NOTICE

@@ -0,0 +1,3 @@
+The Digital Architect of Prestige — premium design skill for coding agents
+Copyright 2026 WizeMe.APP
+Dual-licensed under Apache-2.0 OR MIT at your option.

code_factory_4_design-0.3.0/PKG-INFO

@@ -0,0 +1,154 @@
+Metadata-Version: 2.4
+Name: code-factory-4-design
+Version: 0.3.0
+Summary: The Digital Architect of Prestige — a coding-agent skill for designing premium, high-conversion websites and mobile UIs using applied psychology and trust engineering. Ships a SKILL.md contract plus an executable design-linter that audits generated HTML/CSS against the Halo Effect, Cognitive Fluency, Trust Engineering, Peak-End Rule, and Horn-Effect triggers.
+License-Expression: MIT OR Apache-2.0
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE-APACHE
+License-File: LICENSE-MIT
+License-File: NOTICE
+Provides-Extra: dev
+Requires-Dist: pytest>=8.0; extra == "dev"
+Dynamic: license-file
+
+# The Digital Architect of Prestige 💎
+
+A **coding-agent skill** for designing premium, high-conversion websites and
+mobile UIs — for Claude Code, Codex, OpenCode, Cursor, and any agent that reads
+a project skill file. It moves the agent beyond aesthetics into **applied
+psychology and trust engineering**, and ships an **executable design linter** so
+the agent can *verify* its output against the principles, not just intend them.
+
+> Users judge credibility in ~50ms from visuals alone. If it looks
+> professional, they assume the service is too (the Halo Effect). This skill
+> engineers that halo — and refuses to let a pretty page ship over a broken one.
+
+## Workflow at a glance
+
+```mermaid
+flowchart LR
+    A["User journey or page brief"] --> B["Select workflow and purpose"]
+    B --> C["Build premium UI"]
+    C --> D["Audit five laws"]
+    D --> E["Score bias, CTA, hook, mobile, and purpose fit"]
+    E -->|"fails"| F["Fix ranked recommendations"]
+    F --> D
+    E -->|"passes"| G["Ship UI with design receipt"]
+```
+
+## Install into your agent (any OS)
+
+```bash
+python install.py                 # Windows / macOS / Linux (installs + verifies)
+# then, inside your project:
+prestige install claude           # or: codex · opencode · cursor · generic
+```
+That drops `SKILL.md` where your agent auto-reads it. From then on, any
+"design a landing page / make this premium / build a checkout" request triggers
+the skill.
+
+## What the agent gets
+
+**A design contract (`SKILL.md`)** encoding five laws it applies every time:
+
+1. **The 50ms Halo** — hero section, one bold headline, high-fidelity cover imagery.
+2. **Cognitive Fluency** — Hick's/Miller's Law, extreme whitespace, one goal per section, `line-height ≥ 1.6`.
+3. **Trust Engineering** — security + social-proof cues at decision points, total price incl. fees upfront, guest checkout.
+4. **Peak-End Rule** — micro-interactions at the peaks, inline form validation, a rewarding confirmation ending.
+5. **Horn-Effect Defense** — one Visual DNA in CSS variables, mobile-first stability, no dated cues.
+
+**An executable linter (`prestige audit`)** that scores generated HTML/CSS on all
+five laws and **hard-fails on major functional flaws** — because the skill's one
+non-negotiable rule is *never let aesthetics mask a broken primary action or an
+inaccessible UI.*
+
+```bash
+prestige scaffold site.html        # premium starter (scores 100/100)
+prestige audit site.html           # score the five laws + catch Horn-Effect triggers
+prestige purpose site.html --purpose developer  # opinionated purpose-fit audit
+prestige audit site.html --strict  # exit 1 if failing — use as an agent/CI gate
+```
+
+Example on a deliberately clunky page:
+```
+Overall: 18/100   Verdict: NEEDS WORK
+halo:15  fluency:20  trust:20  peak:20  horn:15
+⛔ HARD FAILURES (fix before shipping — function over aesthetics):
+   MAJOR_FLAW: primary action button has no accessible label/text.
+   ✗ [trust] T_NO_SECURITY: Transactional UI with no security cue near actions.
+   ✗ [horn]  C_NO_VIEWPORT: Missing viewport meta — mobile layout will break.
+```
+
+## The workflow the skill teaches
+
+`Discovery` (audit the journey for trust gaps) → `Build` (apply the five laws)
+→ `Verify` (`prestige audit`, fix flags, then present). The agent treats a
+failing audit like a failing test — which is exactly how "premium" stops being
+a vibe and becomes a checkable property.
+
+## The psychology, sourced
+Halo Effect & the 50ms window · Aesthetic-Usability Effect · Hick's & Miller's
+Law · Fitts's Law · Peak-End Rule · Negativity/Horn Effect. Detailed playbooks
+in `references/` (hero-halo, cognitive-fluency, trust-engineering, peak-end).
+
+## v0.2 — deeper psychology, stricter scoring, five workflows
+
+- **Cognitive Bias Engine** — six research-backed biases (Von Restorff, Anchoring,
+  Social Proof, Loss Aversion, Reciprocity, Cognitive Fluency) with detectors,
+  quality scoring, and precise recommendations.
+- **CTA & Hook Verification** — scores whether your call-to-action and headline
+  are engineered to convert (weak-verb detection, reader-focus, power words,
+  CTA-competition penalty) with exact rewrites.
+- **Strict Mobile Judge** — thumb-reach tap targets, viewport discipline, real
+  breakpoints, legible base text, sticky-CTA guidance — because trust breaks on
+  mobile first.
+- **Five Design-Principle Workflows** — Conversion Architect, Luxury Minimalist,
+  Trust Engineer, Editorial Storyteller, Product-Led Pragmatist. Each is a
+  distinct optimization strategy that re-weights the score.
+- **Precision Scoring** — `prestige score <file> --workflow <key>` fuses laws +
+  biases + CTA/hooks + mobile into a 0–100 conversion-readiness grade with
+  **recommendations ranked by conversion impact**.
+
+```bash
+prestige workflows                          # see the five lenses
+prestige score site.html --workflow trust   # precise score + ranked fixes
+prestige score site.html --strict           # exit 1 if not conversion-ready
+```
+
+Grounded in current research: credibility forms in ~0.05s (Google), trigger-driven
+UX lifts conversion ~34% on average (Nielsen Norman 2024).
+
+## v0.3 - purpose-fit design judgment
+
+Workflows answer **what the page optimizes for**. Purpose lenses answer **what
+psychological job the design must do for this audience and domain**.
+
+Prestige now ships deterministic purpose-fit profiles:
+
+- `developer` - concrete proof, docs, CLI/API clarity, GitHub/demo trust.
+- `healthcare` - calm reassurance, privacy, clinician proof, no miracle hype.
+- `fintech` - security, transparent fees/rates, control, compliance cues.
+- `luxury` - restraint, craft, whitespace, fewer louder elements.
+- `marketplace` - bilateral buyer/seller trust, reviews, protection, comparison.
+- `saas` - product visibility, integrations, ROI, low-friction activation.
+- `editorial` - narrative rhythm, sources, evidence, satisfying end beat.
+
+```bash
+prestige purposes
+prestige purpose site.html --purpose healthcare --strict
+prestige score site.html --workflow trust --purpose healthcare
+```
+
+The purpose gate scores intent clarity, proof fit, visual theme fit, action
+language, and purpose-specific anti-patterns. This makes Prestige more
+opinionated in the right way: not just "pretty," but fit for the decision the
+interface exists to change.
+
+## License
+Dual-licensed **Apache-2.0 OR MIT** — pick whichever your project prefers.
+## Criterion attribution
+
+Prestige 0.2.1 wraps its existing five per-law scores in the factory attribution
+envelope. Run `prestige audit page.html --json` to receive one unit per design
+criterion, its score and threshold, and the unchanged overall verdict.

code_factory_4_design-0.3.0/README.md

@@ -0,0 +1,140 @@
+# The Digital Architect of Prestige 💎
+
+A **coding-agent skill** for designing premium, high-conversion websites and
+mobile UIs — for Claude Code, Codex, OpenCode, Cursor, and any agent that reads
+a project skill file. It moves the agent beyond aesthetics into **applied
+psychology and trust engineering**, and ships an **executable design linter** so
+the agent can *verify* its output against the principles, not just intend them.
+
+> Users judge credibility in ~50ms from visuals alone. If it looks
+> professional, they assume the service is too (the Halo Effect). This skill
+> engineers that halo — and refuses to let a pretty page ship over a broken one.
+
+## Workflow at a glance
+
+```mermaid
+flowchart LR
+    A["User journey or page brief"] --> B["Select workflow and purpose"]
+    B --> C["Build premium UI"]
+    C --> D["Audit five laws"]
+    D --> E["Score bias, CTA, hook, mobile, and purpose fit"]
+    E -->|"fails"| F["Fix ranked recommendations"]
+    F --> D
+    E -->|"passes"| G["Ship UI with design receipt"]
+```
+
+## Install into your agent (any OS)
+
+```bash
+python install.py                 # Windows / macOS / Linux (installs + verifies)
+# then, inside your project:
+prestige install claude           # or: codex · opencode · cursor · generic
+```
+That drops `SKILL.md` where your agent auto-reads it. From then on, any
+"design a landing page / make this premium / build a checkout" request triggers
+the skill.
+
+## What the agent gets
+
+**A design contract (`SKILL.md`)** encoding five laws it applies every time:
+
+1. **The 50ms Halo** — hero section, one bold headline, high-fidelity cover imagery.
+2. **Cognitive Fluency** — Hick's/Miller's Law, extreme whitespace, one goal per section, `line-height ≥ 1.6`.
+3. **Trust Engineering** — security + social-proof cues at decision points, total price incl. fees upfront, guest checkout.
+4. **Peak-End Rule** — micro-interactions at the peaks, inline form validation, a rewarding confirmation ending.
+5. **Horn-Effect Defense** — one Visual DNA in CSS variables, mobile-first stability, no dated cues.
+
+**An executable linter (`prestige audit`)** that scores generated HTML/CSS on all
+five laws and **hard-fails on major functional flaws** — because the skill's one
+non-negotiable rule is *never let aesthetics mask a broken primary action or an
+inaccessible UI.*
+
+```bash
+prestige scaffold site.html        # premium starter (scores 100/100)
+prestige audit site.html           # score the five laws + catch Horn-Effect triggers
+prestige purpose site.html --purpose developer  # opinionated purpose-fit audit
+prestige audit site.html --strict  # exit 1 if failing — use as an agent/CI gate
+```
+
+Example on a deliberately clunky page:
+```
+Overall: 18/100   Verdict: NEEDS WORK
+halo:15  fluency:20  trust:20  peak:20  horn:15
+⛔ HARD FAILURES (fix before shipping — function over aesthetics):
+   MAJOR_FLAW: primary action button has no accessible label/text.
+   ✗ [trust] T_NO_SECURITY: Transactional UI with no security cue near actions.
+   ✗ [horn]  C_NO_VIEWPORT: Missing viewport meta — mobile layout will break.
+```
+
+## The workflow the skill teaches
+
+`Discovery` (audit the journey for trust gaps) → `Build` (apply the five laws)
+→ `Verify` (`prestige audit`, fix flags, then present). The agent treats a
+failing audit like a failing test — which is exactly how "premium" stops being
+a vibe and becomes a checkable property.
+
+## The psychology, sourced
+Halo Effect & the 50ms window · Aesthetic-Usability Effect · Hick's & Miller's
+Law · Fitts's Law · Peak-End Rule · Negativity/Horn Effect. Detailed playbooks
+in `references/` (hero-halo, cognitive-fluency, trust-engineering, peak-end).
+
+## v0.2 — deeper psychology, stricter scoring, five workflows
+
+- **Cognitive Bias Engine** — six research-backed biases (Von Restorff, Anchoring,
+  Social Proof, Loss Aversion, Reciprocity, Cognitive Fluency) with detectors,
+  quality scoring, and precise recommendations.
+- **CTA & Hook Verification** — scores whether your call-to-action and headline
+  are engineered to convert (weak-verb detection, reader-focus, power words,
+  CTA-competition penalty) with exact rewrites.
+- **Strict Mobile Judge** — thumb-reach tap targets, viewport discipline, real
+  breakpoints, legible base text, sticky-CTA guidance — because trust breaks on
+  mobile first.
+- **Five Design-Principle Workflows** — Conversion Architect, Luxury Minimalist,
+  Trust Engineer, Editorial Storyteller, Product-Led Pragmatist. Each is a
+  distinct optimization strategy that re-weights the score.
+- **Precision Scoring** — `prestige score <file> --workflow <key>` fuses laws +
+  biases + CTA/hooks + mobile into a 0–100 conversion-readiness grade with
+  **recommendations ranked by conversion impact**.
+
+```bash
+prestige workflows                          # see the five lenses
+prestige score site.html --workflow trust   # precise score + ranked fixes
+prestige score site.html --strict           # exit 1 if not conversion-ready
+```
+
+Grounded in current research: credibility forms in ~0.05s (Google), trigger-driven
+UX lifts conversion ~34% on average (Nielsen Norman 2024).
+
+## v0.3 - purpose-fit design judgment
+
+Workflows answer **what the page optimizes for**. Purpose lenses answer **what
+psychological job the design must do for this audience and domain**.
+
+Prestige now ships deterministic purpose-fit profiles:
+
+- `developer` - concrete proof, docs, CLI/API clarity, GitHub/demo trust.
+- `healthcare` - calm reassurance, privacy, clinician proof, no miracle hype.
+- `fintech` - security, transparent fees/rates, control, compliance cues.
+- `luxury` - restraint, craft, whitespace, fewer louder elements.
+- `marketplace` - bilateral buyer/seller trust, reviews, protection, comparison.
+- `saas` - product visibility, integrations, ROI, low-friction activation.
+- `editorial` - narrative rhythm, sources, evidence, satisfying end beat.
+
+```bash
+prestige purposes
+prestige purpose site.html --purpose healthcare --strict
+prestige score site.html --workflow trust --purpose healthcare
+```
+
+The purpose gate scores intent clarity, proof fit, visual theme fit, action
+language, and purpose-specific anti-patterns. This makes Prestige more
+opinionated in the right way: not just "pretty," but fit for the decision the
+interface exists to change.
+
+## License
+Dual-licensed **Apache-2.0 OR MIT** — pick whichever your project prefers.
+## Criterion attribution
+
+Prestige 0.2.1 wraps its existing five per-law scores in the factory attribution
+envelope. Run `prestige audit page.html --json` to receive one unit per design
+criterion, its score and threshold, and the unchanged overall verdict.

code_factory_4_design-0.3.0/code_factory_4_design.egg-info/PKG-INFO

@@ -0,0 +1,154 @@
+Metadata-Version: 2.4
+Name: code-factory-4-design
+Version: 0.3.0
+Summary: The Digital Architect of Prestige — a coding-agent skill for designing premium, high-conversion websites and mobile UIs using applied psychology and trust engineering. Ships a SKILL.md contract plus an executable design-linter that audits generated HTML/CSS against the Halo Effect, Cognitive Fluency, Trust Engineering, Peak-End Rule, and Horn-Effect triggers.
+License-Expression: MIT OR Apache-2.0
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE-APACHE
+License-File: LICENSE-MIT
+License-File: NOTICE
+Provides-Extra: dev
+Requires-Dist: pytest>=8.0; extra == "dev"
+Dynamic: license-file
+
+# The Digital Architect of Prestige 💎
+
+A **coding-agent skill** for designing premium, high-conversion websites and
+mobile UIs — for Claude Code, Codex, OpenCode, Cursor, and any agent that reads
+a project skill file. It moves the agent beyond aesthetics into **applied
+psychology and trust engineering**, and ships an **executable design linter** so
+the agent can *verify* its output against the principles, not just intend them.
+
+> Users judge credibility in ~50ms from visuals alone. If it looks
+> professional, they assume the service is too (the Halo Effect). This skill
+> engineers that halo — and refuses to let a pretty page ship over a broken one.
+
+## Workflow at a glance
+
+```mermaid
+flowchart LR
+    A["User journey or page brief"] --> B["Select workflow and purpose"]
+    B --> C["Build premium UI"]
+    C --> D["Audit five laws"]
+    D --> E["Score bias, CTA, hook, mobile, and purpose fit"]
+    E -->|"fails"| F["Fix ranked recommendations"]
+    F --> D
+    E -->|"passes"| G["Ship UI with design receipt"]
+```
+
+## Install into your agent (any OS)
+
+```bash
+python install.py                 # Windows / macOS / Linux (installs + verifies)
+# then, inside your project:
+prestige install claude           # or: codex · opencode · cursor · generic
+```
+That drops `SKILL.md` where your agent auto-reads it. From then on, any
+"design a landing page / make this premium / build a checkout" request triggers
+the skill.
+
+## What the agent gets
+
+**A design contract (`SKILL.md`)** encoding five laws it applies every time:
+
+1. **The 50ms Halo** — hero section, one bold headline, high-fidelity cover imagery.
+2. **Cognitive Fluency** — Hick's/Miller's Law, extreme whitespace, one goal per section, `line-height ≥ 1.6`.
+3. **Trust Engineering** — security + social-proof cues at decision points, total price incl. fees upfront, guest checkout.
+4. **Peak-End Rule** — micro-interactions at the peaks, inline form validation, a rewarding confirmation ending.
+5. **Horn-Effect Defense** — one Visual DNA in CSS variables, mobile-first stability, no dated cues.
+
+**An executable linter (`prestige audit`)** that scores generated HTML/CSS on all
+five laws and **hard-fails on major functional flaws** — because the skill's one
+non-negotiable rule is *never let aesthetics mask a broken primary action or an
+inaccessible UI.*
+
+```bash
+prestige scaffold site.html        # premium starter (scores 100/100)
+prestige audit site.html           # score the five laws + catch Horn-Effect triggers
+prestige purpose site.html --purpose developer  # opinionated purpose-fit audit
+prestige audit site.html --strict  # exit 1 if failing — use as an agent/CI gate
+```
+
+Example on a deliberately clunky page:
+```
+Overall: 18/100   Verdict: NEEDS WORK
+halo:15  fluency:20  trust:20  peak:20  horn:15
+⛔ HARD FAILURES (fix before shipping — function over aesthetics):
+   MAJOR_FLAW: primary action button has no accessible label/text.
+   ✗ [trust] T_NO_SECURITY: Transactional UI with no security cue near actions.
+   ✗ [horn]  C_NO_VIEWPORT: Missing viewport meta — mobile layout will break.
+```
+
+## The workflow the skill teaches
+
+`Discovery` (audit the journey for trust gaps) → `Build` (apply the five laws)
+→ `Verify` (`prestige audit`, fix flags, then present). The agent treats a
+failing audit like a failing test — which is exactly how "premium" stops being
+a vibe and becomes a checkable property.
+
+## The psychology, sourced
+Halo Effect & the 50ms window · Aesthetic-Usability Effect · Hick's & Miller's
+Law · Fitts's Law · Peak-End Rule · Negativity/Horn Effect. Detailed playbooks
+in `references/` (hero-halo, cognitive-fluency, trust-engineering, peak-end).
+
+## v0.2 — deeper psychology, stricter scoring, five workflows
+
+- **Cognitive Bias Engine** — six research-backed biases (Von Restorff, Anchoring,
+  Social Proof, Loss Aversion, Reciprocity, Cognitive Fluency) with detectors,
+  quality scoring, and precise recommendations.
+- **CTA & Hook Verification** — scores whether your call-to-action and headline
+  are engineered to convert (weak-verb detection, reader-focus, power words,
+  CTA-competition penalty) with exact rewrites.
+- **Strict Mobile Judge** — thumb-reach tap targets, viewport discipline, real
+  breakpoints, legible base text, sticky-CTA guidance — because trust breaks on
+  mobile first.
+- **Five Design-Principle Workflows** — Conversion Architect, Luxury Minimalist,
+  Trust Engineer, Editorial Storyteller, Product-Led Pragmatist. Each is a
+  distinct optimization strategy that re-weights the score.
+- **Precision Scoring** — `prestige score <file> --workflow <key>` fuses laws +
+  biases + CTA/hooks + mobile into a 0–100 conversion-readiness grade with
+  **recommendations ranked by conversion impact**.
+
+```bash
+prestige workflows                          # see the five lenses
+prestige score site.html --workflow trust   # precise score + ranked fixes
+prestige score site.html --strict           # exit 1 if not conversion-ready
+```
+
+Grounded in current research: credibility forms in ~0.05s (Google), trigger-driven
+UX lifts conversion ~34% on average (Nielsen Norman 2024).
+
+## v0.3 - purpose-fit design judgment
+
+Workflows answer **what the page optimizes for**. Purpose lenses answer **what
+psychological job the design must do for this audience and domain**.
+
+Prestige now ships deterministic purpose-fit profiles:
+
+- `developer` - concrete proof, docs, CLI/API clarity, GitHub/demo trust.
+- `healthcare` - calm reassurance, privacy, clinician proof, no miracle hype.
+- `fintech` - security, transparent fees/rates, control, compliance cues.
+- `luxury` - restraint, craft, whitespace, fewer louder elements.
+- `marketplace` - bilateral buyer/seller trust, reviews, protection, comparison.
+- `saas` - product visibility, integrations, ROI, low-friction activation.
+- `editorial` - narrative rhythm, sources, evidence, satisfying end beat.
+
+```bash
+prestige purposes
+prestige purpose site.html --purpose healthcare --strict
+prestige score site.html --workflow trust --purpose healthcare
+```
+
+The purpose gate scores intent clarity, proof fit, visual theme fit, action
+language, and purpose-specific anti-patterns. This makes Prestige more
+opinionated in the right way: not just "pretty," but fit for the decision the
+interface exists to change.
+
+## License
+Dual-licensed **Apache-2.0 OR MIT** — pick whichever your project prefers.
+## Criterion attribution
+
+Prestige 0.2.1 wraps its existing five per-law scores in the factory attribution
+envelope. Run `prestige audit page.html --json` to receive one unit per design
+criterion, its score and threshold, and the unchanged overall verdict.

code_factory_4_design-0.3.0/code_factory_4_design.egg-info/SOURCES.txt

@@ -0,0 +1,23 @@
+LICENSE-APACHE
+LICENSE-MIT
+NOTICE
+README.md
+pyproject.toml
+code_factory_4_design.egg-info/PKG-INFO
+code_factory_4_design.egg-info/SOURCES.txt
+code_factory_4_design.egg-info/dependency_links.txt
+code_factory_4_design.egg-info/entry_points.txt
+code_factory_4_design.egg-info/requires.txt
+code_factory_4_design.egg-info/top_level.txt
+prestige_design/__init__.py
+prestige_design/adapters.py
+prestige_design/attribution.py
+prestige_design/audit.py
+prestige_design/biases.py
+prestige_design/cli.py
+prestige_design/cta_hooks.py
+prestige_design/mobile_judge.py
+prestige_design/purpose.py
+prestige_design/score.py
+prestige_design/workflows.py
+tests/test_prestige.py

code_factory_4_design-0.3.0/code_factory_4_design.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

code_factory_4_design-0.3.0/code_factory_4_design.egg-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+prestige = prestige_design.cli:main

code_factory_4_design-0.3.0/code_factory_4_design.egg-info/requires.txt

@@ -0,0 +1,3 @@
+
+[dev]
+pytest>=8.0

code_factory_4_design-0.3.0/code_factory_4_design.egg-info/top_level.txt

@@ -0,0 +1 @@
+prestige_design

code_factory_4_design-0.3.0/prestige_design/__init__.py

@@ -0,0 +1,15 @@
+"""The Digital Architect of Prestige — premium web/mobile design as a skill.
+
+Encodes five design laws as an executable audit so any coding agent (Claude
+Code / Codex / OpenCode / Cursor / …) doesn't just *know* the principles but
+can *verify* its own output before shipping:
+
+  1. The 50ms Halo        — hero section, high-fidelity imagery, instant credibility
+  2. Cognitive Fluency    — Hick's/Miller's Law, whitespace, one goal per section
+  3. Trust Engineering    — security cues + social proof at decision points, price transparency
+  4. Peak-End Rule        — micro-interactions at peaks, a rewarding conclusion
+  5. Horn-Effect Defense  — ruthless consistency, mobile stability, no dated cues
+
+The warning is enforced too: aesthetics never mask a MAJOR functional flaw.
+"""
+__version__ = "0.3.0"

code_factory_4_design-0.3.0/prestige_design/adapters.py

@@ -0,0 +1,28 @@
+"""Install this skill into any coding agent. The SKILL.md is the contract;
+each agent reads it from a different location."""
+from __future__ import annotations
+from pathlib import Path
+
+def skill_text() -> str:
+    return (Path(__file__).resolve().parent/"SKILL.md").read_text()
+
+TARGETS = {
+    "claude":   [".claude/skills/prestige-design/SKILL.md"],
+    "codex":    ["AGENTS.md"],           # appended, not overwritten
+    "opencode": [".opencode/skills/prestige-design.md"],
+    "cursor":   [".cursor/rules/prestige-design.md"],
+    "generic":  ["SKILL-prestige-design.md"],
+}
+
+def install(root: Path, agent: str) -> list[Path]:
+    root = Path(root); created = []
+    key = {"claude-code":"claude"}.get(agent, agent)
+    text = skill_text()
+    for rel in TARGETS.get(key, ["AGENTS.md"]):
+        dst = root/rel; dst.parent.mkdir(parents=True, exist_ok=True)
+        if dst.name == "AGENTS.md" and dst.exists():
+            dst.write_text(dst.read_text().rstrip()+"\n\n"+text)
+        else:
+            dst.write_text(text)
+        created.append(dst)
+    return created

code_factory_4_design-0.3.0/prestige_design/attribution.py

@@ -0,0 +1,51 @@
+"""Attribution wrapper for Prestige's existing per-criterion scores."""
+from dataclasses import asdict, dataclass
+from enum import Enum
+
+
+class FailureClass(str, Enum):
+    AMBIGUOUS_REQUIREMENT = "ambiguous_requirement"
+    UNTYPED_INPUT = "untyped_input"
+    SCOPE_ESCAPE = "scope_escape"
+    INVENTED_PARAM = "invented_param"
+    SIGNATURE_DRIFT = "signature_drift"
+    STUB_UNFILLED = "stub_unfilled"
+    COMPLEXITY_EXCEEDED = "complexity_exceeded"
+    INCONSISTENT_LOGIC = "inconsistent_logic"
+    RUNTIME_CRASH = "runtime_crash"
+    RUNTIME_TIMEOUT = "runtime_timeout"
+    WRONG_OUTPUT = "wrong_output"
+    ACCURACY_REGRESSION = "accuracy_regression"
+    NONDETERMINISM = "nondeterminism"
+    SECURITY_FINDING = "security_finding"
+
+
+@dataclass(frozen=True)
+class UnitResult:
+    unit: str
+    stage: str
+    passed: bool
+    evidence: str
+    failure_class: FailureClass | None = None
+
+
+@dataclass
+class Attribution:
+    stage: str
+    n_checked: int
+    n_passed: int
+    units: list[UnitResult]
+
+    @property
+    def rate(self):
+        return self.n_passed / self.n_checked if self.n_checked else 0.0
+
+    def dominant_failure_class(self):
+        return FailureClass.WRONG_OUTPUT if self.n_passed < self.n_checked else None
+
+    def to_dict(self):
+        value = asdict(self)
+        value["rate"] = self.rate
+        dominant = self.dominant_failure_class()
+        value["dominant_failure_class"] = dominant.value if dominant else None
+        return value