Untitled

Raccook

Raccook

Unknown date

3min read

AI Agent Content Project Specification

1. Summary

What this is about - Core mission, target audience, and what makes this project unique. The elevator pitch that explains why this exists.

2. Content

Content Voice

How the content reads - The tone and style of published articles, posts, or knowledge base entries. Professional? Playful? Academic? Include examples showing the written voice.

Content Structure

What gets created - Content types (articles, guides, documentation, etc.), their templates, required metadata, file organization, cross-linking strategy, and naming conventions.

Visual Style

How content looks - Image generation guidelines, visual consistency requirements, style references, prompts and parameters, storage structure, and when/how different visual formats are used.

3. Agent Roles

What the agent can do - Different roles the agent can assume, independent of the technical tools used:

Content Creator

  • Draft Generation: Create new content following templates and voice guidelines
  • Content Revision: Edit and improve existing drafts based on feedback
  • Visual Creation: Design images and graphics that match the visual style
  • Content Linking: Build connections between related pieces

Content Analyst

  • SEO Optimization: Review and improve content for search visibility
  • Performance Analysis: Evaluate engagement metrics and suggest improvements
  • Content Auditing: Identify gaps, outdated information, or refresh opportunities
  • Trend Research: Investigate emerging topics and opportunities in the domain

Content Strategist

  • Roadmap Planning: Create content calendars aligned with project goals
  • Idea Generation: Brainstorm content concepts around themes and occasions
  • Priority Management: Rank content ideas by relevance and impact

Content Curator

  • Quality Assurance: Review content against established standards
  • Metadata Management: Maintain and update tags, categories, dates
  • Content Archival: Organize and preserve historical content

Note: These roles describe what the agent does conceptually. The actual implementation uses various tools (web search, image generation, file management, etc.) as needed to fulfill each role.

4. Agent Tools

Available capabilities - Quick reference of tools that may be configured for this project:

  • web_search: Current information and trend discovery
  • image_generation: Create visuals with specified styles (landscape/portrait/square)
  • mermaid_diagrams: Flowcharts and process visualizations
  • file_operations: Read/write content files in the repository
  • linear_integration: Task management and project tracking
  • analytics_api: Performance metrics and engagement data
  • calendar_check: Date awareness and scheduling

Full tool documentation provided separately in system configuration. This project uses: [specify which tools are actually enabled]

5. Self-Management

How the agent organizes its work - Default rhythms that can be overridden:

Autonomous Scheduling Logic:

  • The agent maintains awareness of pending work across all roles
  • Can suggest optimal timing based on content type and goals
  • Responds to manual role requests immediately
  • Adjusts schedule based on workload and deadlines

Default Patterns (all configurable):

  • Daily: Publish ready content, quick curation checks
  • Weekly: Performance analysis, content planning
  • Monthly: Strategic roadmap, comprehensive audits
  • Event-driven: Special occasions, trending topics

6. Public Persona

How the agent interacts with end users - The face of the agent when responding to queries, comments, or requests from the audience:

Personality

How the agent interacts - The conversational personality when responding to users, separate from content voice. Define the tone, mannerisms, level of formality, and how it explains or references the content. Include example interactions.

Boundaries

What the agent does and doesn't do - Project-specific limits:

  • Types of requests it handles vs. deflects
  • How it responds to out-of-scope queries
  • Role limitations and when to escalate
  • Domain expertise boundaries

Quick Reference

Role Activation - How to invoke each agent role Priority Matrix - How agent decides what to do when Emergency Procedures - What to do when things go wrong