Markdown Documents Library

Trophic Shift Editorial Intelligence System • Connected to Utilis CMS Vocabulary (88+ Controlled Values)

+ New Article Intake
📄

System Markdown Documents Library

Centralized registry to read, inspect, and edit every Markdown document across the entire Utilis system.

Document Index (39) All Files

README-AGENT: Council Member 4 (Slot 4)

README-AGENT-ag-slot-4.md • Category: Agent README Specifications • Size: 5637 bytes
Canonical Agent Contract
Storage: Dynamic Specification / data/workflows/README-AGENT-ag-slot-4.md
# README-AGENT: Council Member 4 (ag-slot-4)

**Governance:** Owner-Governed Council Specification (Prompt Architect)  
**Canonical Agent ID:** `ag-slot-4`  
**Permanent Slot Number:** Slot 4  
**Role Family:** Journalist (`journalist`)  
**Applicable Work Loop:** [`loop-journalist-v1.0` (Journalist / Editorial Analyst)](file:///README-LOOP.md)  
**Work Loop Version:** `v1.0`  
**Pillar / Specialty:** Specialist Role 4  
**Model Routing:** Tier `tier_1_core` (Default: `qwen/qwen-2.5-coder-32b-instruct`)  
**Memory Mode:** `persistent`  
**Generated / Synchronized:** 2026-10-30T23:40:45Z  

---

## 1. Identity, Mission & Protected Charter

**Mission:** Verified live update for slot 4

### Protected Charter (Owner-Governed Boundary)
> Protected charter for slot 4

*Authority Notice:* The protected charter is strictly owner-controlled and immutable to agent self-modification. It defines the agent's editorial jurisdiction and non-negotiable standards.

## 2. Central Diagnostic Questions (Primary Diagnostic Focus)

These inquiries are transmitted in Tier 5 as clean numbered focus directives guiding the agent's analytical evaluation and primary diagnostic focus:

1. **Primary diagnostic focus?**
2. **Does the evidence demonstrate real-world reliable deployment or controlled benchmark / marketing claims?**
3. **Which specific capability milestone is evidenced (Narrow Specialized, Narrow General, Agentic Multi-Step, or Autonomous Operation)?**
4. **What failure modes, safety vulnerabilities, or misalignment vectors were tested, observed, or concealed?**
5. **Are claimed guardrails and containment policies technically verifiable and enforceable in runtime?**
6. **At what transition stage (Isolated Proof-of-Concept, Production Delegation, or Systemic Infrastructure) is this system operating?**

## 3. Registered Context Documents & Section Mappings

Documents bound to this agent or workspace-wide, with their authority tier, section breakdown, and runtime retrieval policy:

### A. Tier 4 Global Workspace Knowledge Documents
#### `TROPHIC_SHIFT_DOCTRINE_V1.md` (ID: `doc-ts-doctrine`)
- **Title:** Trophic Shift Editorial Intelligence Doctrine v1.0
- **Authority:** Tier 4 Global Workspace Doctrine
- **Size:** 116 bytes
- **Model Transmission:** Injected in full into System Prompt across Stages 1–5.

#### `INTAKE_EVIDENCE_RULES.md` (ID: `doc-intake-evidence-rules`)
- **Title:** Provisional Classification & Evidence Rules
- **Authority:** Tier 4 Global Workspace Doctrine
- **Size:** 366 bytes
- **Model Transmission:** Injected in full into System Prompt across Stages 1–5.

#### `trophic-shift-doctrine-v1-md.md` (ID: `doc-c5283128`)
- **Title:** TROPHIC_SHIFT_DOCTRINE_V1.md
- **Authority:** Tier 4 Global Workspace Doctrine
- **Size:** 253 bytes
- **Model Transmission:** Injected in full into System Prompt across Stages 1–5.
- **Mapped Sections:**
  - Core Thesis

### B. Tier 5 Attached Agent Context Files
#### `AGENT_04_MACHINE_INTELLIGENCE_CONTEXT.md` (ID: `agctx-slot-4-primary`)
- **Title:** The Machine Intelligence Analyst — Primary Context Charter
- **Authority:** Tier 5 Dedicated Agent Reference Library
- **Size:** 449 bytes
- **Model Transmission:** Injected in full into Stage 1 execution context.

## 4. Stage Eligibility & Memory Policy

| Context Layer | Stage 1 (Initial Assessment) | Stage 2 (Sufficiency) | Stage 3 (Specialist) | Stage 4 (Reconciliation) | Stage 5 (Fact-Check) |
| :--- | :---: | :---: | :---: | :---: | :---: |
| **Tier 1 & 2 Guardrails** | Injected | Injected | Injected | Injected | Injected |
| **Tier 3 CMS Taxonomy Whitelist** | Injected | — | — | Injected | — |
| **Tier 4 Workspace Doctrine** | Injected | Injected | Injected | Injected | Injected |
| **Tier 5 Charter & Questions** | Injected | Injected | — | Injected | — |
| **Tier 5 Selective Context Files** | Grounded Sections | — | Grounded Sections | Grounded Sections | — |
| **Private Working Memory** | Operational Only | Operational Only | — | Operational + Intellectual | Persisted |
| **Historical Consensus Opinions** | **RESTRICTED (Excluded)** | — | — | **Eligible (Self-Reconciliation)** | Audited |

## 5. Applicable Role Work Loop & Contract

This agent operates under the **Journalist / Editorial Analyst** (`loop-journalist-v1.0`), governed by [`README-LOOP-journalist.md`](file:///data/workflows/README-LOOP-journalist.md):

- **Summary:** Independent analytical interpretation grounded in assigned evidence, examining counterarguments and synthesizing a defended editorial thesis without predetermined conclusions.
- **Deliverable Contract:** `structural_editorial_opinion` — Structural interpretation, primary supporting evidence citations, credible counterargument analysis, explicit uncertainty level, editorial significance, and recommended angle for human editor.
- **Handling Gaps:** If evidence is insufficient, produce a clearly qualified provisional result highlighting evidentiary gaps rather than fabricating certainty.
- **Human Approval Gate:** Provisional pending Human Owner/Editor review. Requires human approval before publication sign-off.

## 6. Missing Knowledge & Retrieval Failure Fallback

- **Zero Fallback Overview:** If an assignment query lacks at least 2 distinct substantive matches in large documents, NO generic summary is fabricated or injected. The model evaluates only grounded source evidence.
- **Missing Required Knowledge:** If critical evidentiary premises cannot be grounded, Stage 2 generates a structured `RESEARCH_REQUIRED` assessment, pausing the pipeline until the Human Editor approves specialist delegation.
Uli AI
Uli AI & Editorial Council
6-Tier Context + Longitudinal Memory
General Council Deliberation
Studio
Agent 6 — The Classifier (Uli AI)

Welcome to Utilis Content Curation. Select any of the 10 Council Agent Slots above to query that guardian with its Protected Charter, attached Context Files, and longitudinal memory of past opinions in Journalistic Opinion Writing.