PhD-Zero/docs/index.html at main · TenureAI/PhD-Zero · 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
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
<!doctype html>
<html lang="en">
  <head>
    <meta charset="utf-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1" />
    <title>PhD-Zero</title>
    <meta
      name="description"
      content="PhD-Zero is an operating system for the autonomous AI scientist, built as a reusable skill layer for Codex, Claude Code, and future research agents."
    />
    <link rel="icon" href="../assets/phd-zero-mark.svg" type="image/svg+xml" />
    <link rel="stylesheet" href="./styles.css" />
  </head>
  <body>
    <div class="page-shell">
      <header class="hero">
        <nav class="topbar">
          <a class="brand" href="#top">
            <img src="../assets/phd-zero-mark.svg" alt="PhD-Zero logo" />
            <span>PhD-Zero</span>
          </a>
          <div class="topbar-links">
            <a href="https://github.com/TenureAI/PhD-Zero">GitHub</a>
            <a href="../README.md">README</a>
            <a href="../README_zh.md">中文</a>
          </div>
        </nav>

        <section class="hero-grid" id="top">
          <div class="hero-copy">
            <p class="eyebrow">Research Infrastructure for Coding Agents</p>
            <h1>An Operating System for the Autonomous AI Scientist</h1>
            <p class="lede">
              PhD-Zero turns AI research workflows into reusable agent skills. It gives coding agents a
              disciplined operating layer for planning, evidence gathering, experiment execution, memory,
              human review, and paper writing.
            </p>
            <div class="hero-actions">
              <a class="button primary" href="https://github.com/TenureAI/PhD-Zero">View Repository</a>
              <a class="button secondary" href="#quick-start">Quick Start</a>
            </div>
            <ul class="hero-points">
              <li>Built for Codex and Claude Code</li>
              <li>Structured around reusable research skills</li>
              <li>Designed for evidence-backed execution, not one-shot prompting</li>
            </ul>
          </div>

          <div class="hero-card">
            <div class="card-header">Core Skill Stack</div>
            <div class="skill-list">
              <div class="skill-item"><strong>run-governor</strong><span>Execution discipline and safety</span></div>
              <div class="skill-item"><strong>deep-research</strong><span>Literature search and evidence synthesis</span></div>
              <div class="skill-item"><strong>experiment-execution</strong><span>Code runs, debugging, and experiments</span></div>
              <div class="skill-item"><strong>memory-manager</strong><span>Working and reusable research memory</span></div>
              <div class="skill-item"><strong>paper-writing</strong><span>Drafting papers and research artifacts</span></div>
            </div>
          </div>
        </section>
      </header>

      <main>
        <section class="section">
          <div class="section-heading">
            <p class="eyebrow">Why It Exists</p>
            <h2>Strong models still need research discipline</h2>
          </div>
          <div class="feature-grid">
            <article class="feature-card">
              <h3>From prompts to workflows</h3>
              <p>PhD-Zero decomposes research into reusable stages instead of treating it as a single response.</p>
            </article>
            <article class="feature-card">
              <h3>From guesses to evidence</h3>
              <p>Skills are built around search, verification, experiments, and explicit reporting.</p>
            </article>
            <article class="feature-card">
              <h3>From isolated runs to memory</h3>
              <p>Project context and memory skills let agents accumulate reusable research state over time.</p>
            </article>
          </div>
        </section>

        <section class="section">
          <div class="section-heading">
            <p class="eyebrow">How It Works</p>
            <h2>One skill source, multiple agent runtimes</h2>
          </div>
          <div class="stack-layout">
            <div class="stack-column">
              <span class="stack-label">Source of Truth</span>
              <div class="stack-box"><code>.agents/skills/</code></div>
            </div>
            <div class="stack-arrow">→</div>
            <div class="stack-column">
              <span class="stack-label">Codex / Copilot-style agents</span>
              <div class="stack-box"><code>AGENTS.md</code> + workspace skills</div>
            </div>
            <div class="stack-arrow">→</div>
            <div class="stack-column">
              <span class="stack-label">Claude Code</span>
              <div class="stack-box"><code>.claude/skills/</code> mirror layer</div>
            </div>
          </div>
        </section>

        <section class="section" id="quick-start">
          <div class="section-heading">
            <p class="eyebrow">Quick Start</p>
            <h2>Get the skill layer wired correctly</h2>
          </div>
          <div class="quickstart-grid">
            <article class="code-card">
              <h3>1. Clone</h3>
              <pre><code>git clone https://github.com/TenureAI/PhD-Zero.git
cd PhD-Zero</code></pre>
            </article>
            <article class="code-card">
              <h3>2. Inspect skills</h3>
              <pre><code>find .agents/skills -mindepth 1 -maxdepth 1 -type d
find .claude/skills -mindepth 1 -maxdepth 1 -type l</code></pre>
            </article>
            <article class="code-card">
              <h3>3. Use with your agent</h3>
              <pre><code># Codex / Copilot-style agents
cat AGENTS.md

# Claude Code discovers mirrored skills
find .claude/skills -mindepth 1 -maxdepth 1 -type l</code></pre>
            </article>
          </div>
          <p class="section-note">
            If the skill names listed under <code>.agents/skills</code> and <code>.claude/skills</code> match,
            the shared skill layer is wired correctly.
          </p>
        </section>

        <section class="section">
          <div class="section-heading">
            <p class="eyebrow">Roadmap</p>
            <h2>From research intern to autonomous scientist</h2>
          </div>
          <div class="roadmap">
            <article class="roadmap-step">
              <span class="roadmap-phase">Phase 1</span>
              <h3>AI Research Intern</h3>
              <p>Scoped literature investigation, experiment planning, execution, debugging, and report drafting.</p>
            </article>
            <article class="roadmap-step">
              <span class="roadmap-phase">Phase 2</span>
              <h3>AI Research Collaborator</h3>
              <p>Persistent project context, reusable memory, better iteration loops, and stronger checkpoints.</p>
            </article>
            <article class="roadmap-step">
              <span class="roadmap-phase">Phase 3</span>
              <h3>Autonomous AI Scientist</h3>
              <p>Independent hypothesis generation, self-directed experimentation, and long-horizon research execution.</p>
            </article>
          </div>
        </section>
      </main>
    </div>
  </body>
</html>