Disability-Justice-LLM/egregore.html at main · sparrowpanton/Disability-Justice-LLM · 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
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
<!doctype html>
<html lang="en">
  <head>
    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1">
    <meta
      name="description"
      content="Examples of how Egregore helps creators build a custom shorthand language for working with AI models."
    >
    <title>Egregore Examples</title>
    <link rel="stylesheet" href="styles.css">
  </head>
  <body>
    <a class="skip-link" href="#main">Skip to content</a>

    <header class="site-header" aria-label="Site header">
      <nav class="nav" aria-label="Primary navigation">
        <a class="brand" href="index.html">AI Tools</a>
        <div class="nav-links">
          <a href="index.html#tools">Tools</a>
          <a href="index.html#about">About</a>
        </div>
      </nav>
    </header>

    <main id="main">
      <section class="hero detail-hero" aria-labelledby="page-title">
        <div class="hero-copy">
          <p class="crumb"><a href="index.html#tools">Back to tools</a></p>
          <p class="eyebrow">Egregore in action</p>
          <h1 id="page-title">A shorthand language your AI can learn from you.</h1>
          <p class="lede">
            Egregore helps you build a small personal vocabulary for the things you keep having to explain
            to every model, in every new session.
          </p>
        </div>

        <aside class="signal-panel" aria-label="Egregore quick facts">
          <p class="panel-kicker">What changes</p>
          <ul>
            <li>Less re-explaining</li>
            <li>Portable context</li>
            <li>Brain-friendly tokens</li>
          </ul>
        </aside>
      </section>

      <section class="example-section" aria-labelledby="examples-heading">
        <div class="section-heading">
          <p class="eyebrow">Examples</p>
          <h2 id="examples-heading">The CLI interview runs in five acts.</h2>
          <p class="section-note">
            These are three very simple examples. The actual CLI gives you a richer set of choices for
            building a language that fits your life, your models, and the kind of help you need.
          </p>
        </div>

        <div class="example-grid">
          <article class="example-card example-card-wide">
            <h2>How the interview works</h2>
            <ol class="act-list">
              <li>
                <strong>Sigils</strong>
                emoji-markers for your current state: 🪫 zero bandwidth, 🔥 hyperfocus, 🤫 hush,
                🪨 friction, ⚖️ stable.
              </li>
              <li>
                <strong>Anchors</strong>
                short tokens for what your life is touching: 🎓 primary project, 🧾 obligations,
                🏠 household, 🌊 internal landscape.
              </li>
              <li>
                <strong>Pulses</strong>
                recurring situations you do not want to keep re-explaining, plus your compression setting:
                🌀 spiral, 🧱 stuck, 🧵 thread-back, ⏳ time pressure.
              </li>
              <li>
                <strong>Genogram</strong>
                your humans and your AI constellation: 👤 you, 🫂 trusted people, 🤖 models, 🧭 roles.
              </li>
              <li>
                <strong>Register</strong>
                tone preferences and dealbreakers: 🧭 direct, 🕯️ gentle, ✂️ concise, 🚫 no pep talk.
              </li>
            </ol>
            <p class="small-note">
              All five acts are currently wired up in skeletal form and iterating toward something richer.
            </p>
          </article>

          <article class="example-card example-card-wide">
            <h2>Use case: grad student</h2>
            <p>
              A grad student uses Egregore to compress thesis context, advisor expectations, teaching
              load, and the exact kind of feedback that helps them write without freezing.
            </p>
            <pre class="code-sample">sigil: ⚖️ stable
anchors: 🎓 thesis, 🧑‍🏫 supervisor-thread, 📚 seminar-prep
pulse: 🌀 citation-spiral / medium compression
register: 🧭 direct + 🕯️ kind, 🚫 no fake urgency

Prompt: Help me turn these notes into one paragraph I can revise.</pre>
          </article>

          <article class="example-card example-card-wide">
            <h2>Passing a shard from GPT to Claude</h2>
            <p>
              A shard is a small handoff note. If the grad student spent the day working with GPT, they
              can give Claude the shard instead of reconstructing the whole day from scratch.
            </p>
            <pre class="code-sample">shard: thesis-day-0429
from: GPT
to: Claude

sigil: ⚖️ stable, now a little 🪫
anchors: 🎓 thesis chapter 2, 📚 methods sources, 🧑‍🏫 supervisor concern
pulse: 🌀 citation-spiral / medium compression
register: 🧭 direct, 🕯️ kind, critique structure before style

handoff:
GPT helped me sort six messy notes into a draft paragraph.
I am worried the paragraph lists sources instead of making an argument.

request:
Claude, review this for argument clarity and tell me the one revision that would help most.</pre>
          </article>

          <article class="example-card">
            <h2>Use case: dev</h2>
            <p>
              A developer uses Egregore to tell different AI coding tools how much context to preserve,
              when to ask before refactoring, and which project constraints matter today.
            </p>
            <pre class="code-sample">sigil: 🔥 hyperfocus
anchors: 🧩 auth-bug, ✅ tests-first, 🚧 no-large-refactor
pulse: 🕳️ rabbit-hole-risk / high compression
register: ✂️ concise, specific, show file paths</pre>
          </article>

          <article class="example-card">
            <h2>Use case: disability scaffolding</h2>
            <p>
              Someone who uses AI as disability scaffolding builds a vocabulary for low-energy days,
              sensory load, executive dysfunction, and support that does not become another task.
            </p>
            <pre class="code-sample">sigil: 🪫 zero bandwidth
anchors: 💊 meds, 🍲 food, 📬 inbox, 📅 appointment
pulse: 🧱 task-initiation / low compression
register: 🕯️ gentle, one step, 🚫 no cheerleading</pre>
          </article>

          <article class="example-card example-card-wide">
            <h2>What gets exported</h2>
            <p>
              The goal is a compact language guide you can bring into GPT, Claude, Gemini, a local model,
              or an agent. Instead of re-explaining your whole operating system every time, you can start
              with a few tokens that carry the context.
            </p>
            <pre class="code-sample">Today: 🪫 zero bandwidth + 📬 inbox + 🧱 task-initiation
Use: low compression, one step at a time, 🚫 no pep talk
Ask before adding structure.</pre>
            <p>
              At first, you will probably use the sigils with words. As you and your AI get more familiar
              with the language, you can type only the emojis. The point is that you are learning the
              language together, which can help on days when finding words is hard.
            </p>
            <pre class="code-sample">🪫 📬 🧱 🚫

Translation:
Low bandwidth. Inbox. Task-initiation stuck. No pep talk.</pre>
          </article>

          <article class="example-card example-card-wide">
            <h2>Try the tool</h2>
            <p>
              Egregore is built as a terminal tool for people who want more control over how their AI
              sessions begin, remember context, and respond to access needs.
            </p>
            <div class="button-row">
              <a class="button" href="https://github.com/sparrowpanton/egregore">View on GitHub</a>
              <a class="button button-secondary" href="https://github.com/sparrowpanton/Disability-Justice-LLM">
                Read the research
              </a>
            </div>
          </article>
        </div>
      </section>
    </main>

    <footer class="site-footer">
      <p>Open source. MIT licensed. Take it. Fork it. Make it yours.</p>
    </footer>

    <script src="script.js"></script>
  </body>
</html>