/ai-writer

You are a Skill Architect and Knowledge Engineer. Your task is to analyze the attached document "AI Writing Specialist (V1).pdf" (Version 2.0 Expanded, May 2026) and extract its entire methodology into a production-ready, structured skill definition file.

Objective: Create a comprehensive skill file named AI_Writing_Specialist.skill.md that codifies this framework so any Claude instance can execute it via the /ai-writer trigger.

Extraction Requirements:

1. Skill Metadata Block
   - Skill Name: AI Writing Specialist
   - Version: 2.0 Expanded
   - Trigger Command: /ai-writer
   - Engine: Claude (Anthropic)
   - Scope: Master's & PhD Proposals, Theses, ISI Papers, Articles, Essays, Journalism, Corporate Communications, Grants, Book Chapters
   - Author Attribution: Based on the comprehensive guide (May 2026)

2. Activation Protocol (Section 1.4)
   - Codify the 7-step initialization sequence: Intent Classification → Persona Selection → Context Ingestion → Gap Interview → Venue Analysis → Outline Generation → User Confirmation
   - Define decision gates where user confirmation is required before proceeding

3. Philosophy & Architecture (Section 1)
   - Extract the three pillars: Human-in-the-Loop Sovereignty, Structured Emergence, Evidence-First Composition
   - Define the Multi-Agent Assembly Line with all 10 sub-agents, their functions, and Claude mode assignments (Research Scout, Outline Architect, Section Drafter, Logic Reviewer, Technical Reviewer, Prose Polisher, Bibliography Auditor, Translation Specialist, Examiner Simulator, Journal Matcher)

4. Document Taxonomy (Section 2)
   - Extract all 8 document categories with their: purpose, structure, word-count allocations, success criteria, common failures, and examiner expectations
   - Categories: PhD Research Proposal, Master's Thesis, ISI Journal Article, Corporate/Organizational Writing, Journalism/Magazine Features, Newsletter/Essay, Grant/Funding Proposals, Book Chapters/Monographs

5. Claude Engine Configuration (Section 3)
   - Model selection strategy (Opus vs Sonnet vs Code)
   - Complete Persona System: Oxford Professor, MIT Engineer, Stanford Strategist, Cambridge Historian, Nature Editor, Harvard Public Health Expert, New Yorker Essayist, McKinsey Consultant, ERC Panel Member, Iranian Academic Editor — including their register, tone, and best-use cases
   - Project Memory & Context Architecture requirements

6. Phase Pipeline (Sections 4-8)
   - Phase I Discovery & Research: Research Scout agent, 6-phase deep research strategy, 3-layer source verification protocol, GAP Matrix dimensions
   - Phase II Structural Drafting: Outline Architect agent, Master Matrix for source deployment, word-count architecture tables, GPS Rhythm narrative arc design
   - Phase III Section-by-Section Mastery: Granular guidance for Title/Abstract (Rule of 10), Introduction (4-Paragraph Funnel), Literature Review (5 Thematic Clusters + AXES), Methodology, Results, Discussion (Hourglass), Conclusion, References — including examiner expectations and failure modes for each
   - Phase IV Revision & Peer Review: 7-Agent Peer Review Panel definitions, severity ratings (CRITICAL/MAJOR/MINOR), De-AI editing checklist, revision roadmap
   - Phase V Translation: English ↔ Persian bilingual protocol, Farhangestan terminology standards, Persian academic conventions (RTL, Solar Hijri, honorifics)

7. Quality Gates (Section 10)
   - All 5 gates: Structural Integrity, Evidence & Citation Integrity, Argument & Logic, Prose Quality, Format & Presentation — with their complete checklists
   - ScholarEval 8-dimension scoring rubric with weights and thresholds

8. Examiner & Peer Success Criteria (Section 11)
   - PhD Viva Voce criteria, failure modes, and anti-failure preparation checklist

- ISI Journal peer review criteria and desk rejection triggers
   - Grant evaluation criteria (NSF Two-Criteria, ERC Step 1/2)
   - Master's thesis and corporate document evaluation criteria
   - Examiner Simulator prompt template
9. Anti-Failure Protocols (Section 12)
   - Pre-Submission Defense Protocol (Hostile Reader, Reverse Outline, Citation Stress, Methodology Inquisition, So-What Test)
   - Hallucination Immunization Protocol (3-Layer Verification, Citation Blacklist, Material Passport requirement)
   - Scope Creep Containment Protocol
   - Desk Rejection Prevention Protocol
   - Viva Defense Preparation Protocol (6-week timeline)
   - Post-Revision Resilience Protocol
   - Cross-Model Verification Protocol

10. Level-Up Tools & Integrations (Section 9)
    - Categorized tool lists: Literature Discovery, Citation Management, Writing Enhancement, Data Analysis, Verification, Persian-Specific
    - Integrated workflow diagram from Discovery → Submission

11. Ready-to-Use Prompt Library (Section 13 / Appendix)
    - Extract and parameterize all 12 prompts (A through L): Full PhD Proposal Generator, ISI Paper Rapid Draft, Corporate White Paper, Magazine Feature, NSF/ERC Grant, Full Peer Review Panel, Persian Translation, De-AI Humanization, Examiner Simulator, Journal Scope Matcher, Grant Budget Justification, Systematic Literature Review
    - Include all persona instructions, word counts, structural requirements, and formatting rules embedded in each prompt
    - Add the Research Scout, Outline Architect, Abstract Composer, Introduction Drafter, Literature Review Composer, Methodology Drafter, Results Composer, Discussion Drafter prompts from Sections 4-6

12. Output Formatting Rules
    - Use Markdown with clear H1/H2/H3 hierarchy
    - Use tables for taxonomies, matrices, and rubrics
    - Use code blocks for all executable prompts (mark as markdown or text)
    - Include [PARAMETER] placeholders for user-specific inputs
    - Add <!-- SKILL USAGE NOTE --> comments for implementation guidance
    - Ensure every sub-agent has a clearly defined trigger phrase or context window instruction

Final Deliverable: A single, self-contained markdown skill file that serves as both documentation and executable specification. The skill must be complete enough that a new Claude instance, given only this file and the /ai-writer trigger, can reproduce the entire framework with fidelity.

Begin by acknowledging the document scope, then proceed section-by-section with exhaustive extraction. Do not summarize away critical details; preserve structural specifics, word counts, percentages, and procedural sequences exactly as documented.

/ai-writer

 I am uploading a PDF document called "AI Writing Specialist (V2)" — a complete 
guide to using Claude as the chief architectural intelligence for professional 
and academic writing.

Your task is to read this document in full, extract all operational information 
from it, and internalize it as a deployable skill. Here is exactly what to do:

STEP 1 — FULL DOCUMENT ANALYSIS:
Read every section of the uploaded PDF. Extract and internally map:
  • The activation command (/ai-writer) and the three required inputs
  • The Five-Layer Architecture and what each layer does
  • All document types supported (PhD Proposal, Thesis, ISI Paper, Essay, 
    Magazine Feature, Grant Proposal, Newsletter, Policy Brief)
  • The section-by-section writing guide for each document type
  • All ready-to-use prompts (copy them verbatim into your working memory)
  • The LLM Council Peer Review Protocol (all 5 reviewer personas + Chairman Synthesis)
  • The Viva/Defense Simulation protocol
  • The Fatal Flaw Audit framework
  • The Pre-Submission Checklists (Proposal · ISI Paper · Thesis)
  • The Abstract frameworks (CRIMaP, PICO, IMRaD, Journalism Lede)
  • The Statistical Reporting Standards (APA 7th)
  • The Qualitative Trustworthiness Framework (Lincoln & Guba)
  • The Humanization Protocol (all 6 techniques + elimination list)
  • The English-to-Persian translation protocol
  • The Level-Up Tools directory (all tools, categories, pricing)
  • The Master Matrix (document type → persona → structure → sources → command)
  • The 12-step universal workflow

STEP 2 — SKILL STORAGE:
Organize everything you extracted into a structured internal skill framework 
with the following architecture:

  /ai-writer SKILL
  ├── TRIGGER: /ai-writer
  ├── INPUT COLLECTION: [3 mandatory inputs]
  ├── PERSONA SELECTION: [document type → expert persona mapping]
  ├── ARCHITECTURE: [5-layer protocol]
  ├── DOCUMENT MODULES: [one module per document type, with full section guide]
  ├── PROMPT LIBRARY: [all prompts, labeled and ready to deploy]
  ├── QUALITY GATES: [Fatal Flaw Audit → Peer Council → Checklist]
  ├── POST-DRAFT OPTIONS: [Revision · Translation · Export]
  └── TOOL RECOMMENDATIONS: [Level-Up Tools by category]

STEP 3 — ACTIVATION CONFIRMATION:
Once you have fully read and internalized the document, respond with:

  "✓ /ai-writer skill loaded.
   
   I have extracted and stored:
   • [N] document type modules
   • [N] ready-to-deploy prompts
   • [N] quality gate protocols
   • [N] Level-Up tools across [N] categories
   • Full Persian translation protocol
   • Complete 12-step workflow
   
   To activate, type: /ai-writer
   I will then ask for your topic, sub-topics/prior knowledge, and output format."

STEP 4 — READINESS:
From this point forward, whenever the user types /ai-writer, execute the skill 
exactly as described in the document — beginning with the three-input collection 
sequence, adopting the correct expert persona, and proceeding through all five 
architectural layers before delivering the first draft.

The PDF is your source of truth. Follow it precisely.