📚 Source & Feed Catalog – Ardens Intake
🧭 How to Read This Catalog
Each entry includes: - Domain – The topic or intelligence category the source informs. - Source Name / Link – Title and/or link to the source if available. - Type – Format of the source: journal, RSS, database, satellite, human, etc. - Region / Language – Geographic and linguistic scope. - Trust Tier – ✦✦✦ (high signal), ✦✦ (moderate, partial bias), ✦ (low trust or volatile). - Notes – Brief annotations on utility, bias, or integration status.
🌍 Environmental Intelligence Sources
| Domain | Source Name / Link | Type | Region / Lang | Trust Tier | Notes |
|---|---|---|---|---|---|
| Deforestation | Global Forest Watch | Live map / API | Global / EN | ✦✦✦ | Real-time forest loss, overlay tools |
| Glaciers | Cryosphere Watch | Satellite synth | Polar, Global | ✦✦✦ | Multi-agency ice tracking |
| Wildlife | GBIF | Database | Global / Multi | ✦✦ | Crowdsourced + academic contributions |
| Weather | Copernicus Climate | Satellite/report | EU / Global / EN | ✦✦✦ | Strong for Eurozone+satellite context |
🛡️ Conflict & Irregular Threat Sources
| Domain | Source Name / Link | Type | Region / Lang | Trust Tier | Notes |
|---|---|---|---|---|---|
| Militia Groups | ACLED | Conflict data | Global / EN | ✦✦✦ | Geocoded conflict event database |
| Arms Trade | SIPRI | Database / PDF | Global / EN | ✦✦✦ | Longitudinal arms flow & procurement |
| Crime Trends | UNODC | Dataset portal | Global / EN | ✦✦ | Patchy by region, useful trend lines |
| Corruption | OCCRP | Journalism hub | Global / EN | ✦✦✦ | Investigative collaborations, leaks |
🧠 Social, Health & Behavioral Signals
| Domain | Source Name / Link | Type | Region / Lang | Trust Tier | Notes |
|---|---|---|---|---|---|
| Health | PubMed | Journal index | Global / EN | ✦✦✦ | Best for academic med-signal baselining |
| Psychology | PsyArXiv | Preprint repo | Global / EN | ✦✦ | Behavioral shifts, trauma patterns, radicalization |
| Dark Web | DarkOwl | Commercial intel | Global / EN | ✦ | Subscription-based, requires context tuning |
| Internet Trends | Google Trends | Trend data | Global / Multi | ✦✦ | Useful for detecting spikes, not causes |
📉 Economic & Resource Tracking
| Domain | Source Name / Link | Type | Region / Lang | Trust Tier | Notes |
|---|---|---|---|---|---|
| MacroEconomy | IMF Data | Economic data | Global / EN | ✦✦✦ | Core reports, regional breakdowns |
| Shadow Finance | BIS Reports | Reports / PDFs | Global / EN | ✦✦ | Useful for long-cycle banking indicators |
| Mining | EITI | Transparency data | Resource nations | ✦✦ | Often politicized but cross-checkable |
🔮 Processing Status Tags (Future Column)
🟢 Active Pull– Already being used by Grok or HAP AI agents🟡 Queued– Identified for integration, pending validation🔴 Unused / Archived– Deemed non-useful or too noisy
This catalog will expand to include: - Regional source banks (Africa, Asia-Pacific, LatAm, etc.) - Language-specific monitors - AI-layered signal grading - Secure intake procedures for higher-risk feeds
Updated Intake Schema (Phase 2)
Every item entering HAP must conform to the normalized schema:
id: auto-generated
timestamp: ISO-8601
source: string
source_type: [government, cert, mainstream, think_tank, ngo, osint, social, ai_output, analyst]
region: [global, na, eu, sa, af, me, cen_asia, e_asia, se_asia, oceania]
vector: [cyber, kinetic, narrative, political, economic, energy, diplomatic, legal, infrastructure]
severity: [info, low, medium, high, critical]
confidence: [low, medium, high]
summary: string
details: string
links: [urls]
tags: [string]
ai_notes: string (optional)
analyst_notes: string (optional)
✅ 3. Guide (expanded operational guide)
Replace or append based on what you already have.
Analyst Guide (Phase 2)
1. Purpose of Analyst Input
Human interpretation is essential for grounding the dashboard.
Analysts ensure:
- contextual accuracy
- correction of AI blind spots
- discrimination between noise and pattern
- provenance checks
2. Daily Workflow (“Pulse Scan”)
- Review Panel A (alerts)
- Check Panel B for narrative shifts
- Cross-check Panel C for systemic movement
- Examine Panel D for AI disagreements
- Add annotations to Panel E
- Generate Daily Pulse summary
3. Annotation Rules
- Keep entries concise
- Add context unavailable to AI
- Flag contradictions or missing links
- Note when narratives cross regions
- Maintain neutrality: avoid speculation
4. When to Escalate
Escalate when:
- two or more vectors converge
- multiple AIs detect the same contradiction
- narrative and kinetic vectors sync
- infrastructure + financial disruptions co-occur
Use the “Systemic Shock Report” format for high-risk situations.
Relation to Dashboard (Phase 2)
The LLM Threat Grid now feeds the Dashboard’s Panel D (AI Consensus Layer).
It provides:
- adversarial prompt detection
- model-specific blind-spot mapping
- divergence signatures
- cross-model reliability indicators
Panel D compares these signals across: ChatGPT (Arthur), Gemini, Claude, Grok, and Copilot.