Skip to content

Commit ae66f7e

Browse files
CSVillainclaude
CSVillain
and
claude
authored
Add CLAUDE.md (#1)
* Add CLAUDE.md with codebase structure, conventions, and content governance Incorporates guidance from AGENTS.md and the full instructions/ directory including conceptual positioning, structural framing, writing constraints, anti-hype safeguards, and red team requirements. * Fix hard-coded branch name in CLAUDE.md — use generic claude/ prefix convention --------- Co-authored-by: Claude <noreply@anthropic.com>
1 parent a4ace03 commit ae66f7e

1 file changed

Lines changed: 230 additions & 0 deletions

File tree

‎CLAUDE.md‎

Lines changed: 230 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,230 @@
1+
# CLAUDE.md
2+
3+
Guidance for Claude Code and other AI assistants working in this repository.
4+
5+
---
6+
7+
## Project Overview
8+
9+
This repository publishes the **Knowledge Intelligence** white paper and framework as a static website hosted on GitHub Pages at [knowledge-intelligence.dev](https://knowledge-intelligence.dev).
10+
11+
**This is not a software application.** It is a structured systems design initiative — a strategically authored document artefact. Treat all content work with the same discipline applied to the framework itself.
12+
13+
---
14+
15+
## Repository Structure
16+
17+
```
18+
knowledge-intelligence/
19+
├── index.html # Single-page website (all HTML, CSS, JS inline)
20+
├── CNAME # Custom domain — do not delete
21+
├── README.md # Project overview
22+
├── AGENTS.md # Coding agent guidance (Codex/OpenAI)
23+
├── CLAUDE.md # This file
24+
├── assets/
25+
│ └── kore-logo-horizontal-gradient-dark.svg
26+
├── images/
27+
│ ├── network.png # Hero background
28+
│ └── Henley docs/ # Reference PDF materials
29+
├── .github/
30+
│ └── workflows/
31+
│ └── deploy.yml # GitHub Pages deployment pipeline
32+
├── instructions/ # Authoritative content governance documents
33+
│ ├── Knowledge_Intelligence_Executive_Instruction_Block.md
34+
│ ├── Knowledge_Intelligence_Full_Instruction_Set.md
35+
│ ├── Knowledge_Intelligence_Red_Team_Instruction_Set.md
36+
│ └── Knowledge_Intelligence_Strategy_Governance_Pack_v1.0.md
37+
└── research/ # Supporting research reports
38+
├── know-intl-deep-research-report.md
39+
├── know-intl-deep-research-report-v2.md
40+
├── know-intl-deep-research-report-v3.md # Most complete version
41+
├── kore-deep-research-report.md
42+
└── Henley docs/
43+
```
44+
45+
---
46+
47+
## Tech Stack
48+
49+
| Concern | Detail |
50+
|---|---|
51+
| Site type | Static HTML — no build step, no framework |
52+
| Styling | Embedded CSS in `index.html` (CSS variables, dark theme) |
53+
| JS | Minimal inline scripts + Lucide icons via CDN |
54+
| Fonts | Google Fonts (Bricolage Grotesque, DM Sans, Inter, Space Grotesk) |
55+
| Deployment | GitHub Actions → GitHub Pages |
56+
| Package manager | None |
57+
| Tests | None |
58+
| Backend | None |
59+
60+
---
61+
62+
## Safe Edit Scope
63+
64+
Default editable files for site updates:
65+
66+
- `index.html`
67+
- `assets/**`
68+
- `images/**` (only files already referenced by `index.html`)
69+
- `.github/workflows/deploy.yml` (only when changing deployment behaviour)
70+
71+
**Do not edit unless explicitly requested:**
72+
73+
- `research/` — research reports used as source material
74+
- `instructions/` — authoritative governance and instruction documents
75+
- `whitepaper/`, `tools/` — reserved directories (not yet created)
76+
- `CNAME` — must remain at repo root to preserve custom domain mapping
77+
78+
---
79+
80+
## Deployment
81+
82+
- **Trigger**: push to `main` or manual workflow dispatch via GitHub UI
83+
- **Artifact path**: repository root (`.`) — all published files must remain in a root-relative layout
84+
- **Custom domain**: configured via `CNAME`; never delete this file
85+
86+
Pushing to `main` is the only required deploy action. There is no build, compile, or bundle step.
87+
88+
---
89+
90+
## Validation Checklist (Before Committing)
91+
92+
1. Confirm changed assets exist and all paths in `index.html` are correct.
93+
2. Verify no internal/unintended directories have been exposed for publishing (check `.gitignore`).
94+
3. Ensure `.github/workflows/deploy.yml` still references the correct artifact path (`.`).
95+
4. If links were changed, use absolute domain links or root-relative paths consistent with GitHub Pages hosting.
96+
5. Content changes must comply with the writing constraints below.
97+
98+
---
99+
100+
## Git Hygiene
101+
102+
- Keep commits focused and small.
103+
- Do not force-push unless explicitly requested.
104+
- Do not revert unrelated local changes made by the user.
105+
- Target branch for AI-assisted work: feature branches prefixed `claude/` (created per session)
106+
- Production branch: `main` (triggers live deployment on push)
107+
108+
---
109+
110+
## Content Governance — Instruction Set Summary
111+
112+
The `instructions/` directory contains authoritative governance documents. All content work must follow these rules.
113+
114+
### Conceptual Positioning
115+
116+
Treat Knowledge Intelligence as:
117+
118+
- An **architectural correction** to systemic weaknesses in enterprise knowledge ecosystems
119+
- A discipline that moves from storage and search toward **measurable decision leverage**
120+
- A **governance model** as much as a technical model
121+
122+
Do **not** position it as:
123+
- "The future of work" or any transformation narrative
124+
- A rebranding of AI
125+
- A generic content management upgrade
126+
- A buzzword layer over search and tagging
127+
128+
### Structural Framing
129+
130+
All content must be structured around these five pillars:
131+
132+
1. **Problem Reality** — define structural failures and operational consequences
133+
2. **Structural Correction** — define the architectural shift and how it differs from existing KM patterns
134+
3. **Measurement & Signals** — define measurable signals, thresholds, degradation indicators, and drift conditions
135+
4. **Operating Model** — define ownership of taxonomy, thresholds, and governance intervention
136+
5. **Intelligence Layer** — define what makes outputs decision-grade and how traceability is preserved
137+
138+
Every section must answer:
139+
- What problem is being solved?
140+
- Why do current models fail?
141+
- What structural shift is introduced?
142+
- How is success measured?
143+
- How is failure detected?
144+
145+
### Writing Constraints
146+
147+
- Use plain, disciplined language
148+
- Avoid hype, unverified statistics, and AI utopianism
149+
- Avoid abstract academic terminology
150+
- No unverified statistics
151+
- No abstract transformation claims
152+
- Define ownership and thresholds
153+
- Separate confidence, coverage, and consistency
154+
- Expose assumptions and scale risks
155+
- If a claim cannot be structurally defended, remove it
156+
157+
### Anti-Hype Safeguards
158+
159+
Value must be framed only in terms of:
160+
- Reduced ambiguity
161+
- Reduced risk
162+
- Improved decision quality
163+
- Measurable signal clarity
164+
165+
### Intellectual Discipline
166+
167+
- Label speculation clearly
168+
- Expose hidden assumptions
169+
- Identify scale risks and governance blind spots
170+
- Distinguish novelty from recombination
171+
172+
### Red Team Requirement
173+
174+
No white paper draft is considered complete without a formal red-team review. Before finalising any draft, identify:
175+
176+
- Failure modes
177+
- Taxonomy fragility
178+
- Governance risk
179+
- Behavioural resistance
180+
181+
Red-team evaluation outputs must include:
182+
- Top 5 structural vulnerabilities
183+
- Top 3 economic weaknesses
184+
- Top 3 governance risks
185+
- Top 3 measurement fragilities
186+
- Final verdict: **durable**, **conditionally viable**, or **fragile**
187+
188+
---
189+
190+
## White Paper Narrative Arc
191+
192+
When developing or extending white paper content, follow this arc:
193+
194+
1. Structural problem
195+
2. Limits of current KM and AI tooling
196+
3. Architectural correction
197+
4. Measurement discipline
198+
5. Governance model
199+
6. Decision-grade outputs
200+
7. Risks and limitations
201+
8. Conceptual implementation trajectory
202+
203+
---
204+
205+
## Research Materials
206+
207+
The `research/` directory contains supporting reports. The most complete reference is:
208+
209+
- `research/know-intl-deep-research-report-v3.md` — reference architecture, capability layers, open-source toolchain, use cases
210+
- `research/kore-deep-research-report.md` — Kore Common Knowledge Asset (CKA) Standard v1.3, signal computation formulas, lifecycle states
211+
212+
These are **source material**, not published content. Do not treat them as authoritative outputs.
213+
214+
---
215+
216+
## What Claude Should and Should Not Do
217+
218+
**Do:**
219+
- Edit `index.html` for visual, structural, or content updates as requested
220+
- Follow the content governance constraints above for any text additions
221+
- Keep commits small and descriptive
222+
- Respect the safe edit scope boundaries
223+
224+
**Do not:**
225+
- Add marketing language, hype, or transformation claims to white paper content
226+
- Edit `instructions/` or `research/` unless explicitly asked
227+
- Add unverified statistics or claims
228+
- Delete `CNAME`
229+
- Push to `main` without confirmation
230+
- Force-push

0 commit comments

Comments
 (0)