Skill: Agent Task Generator

Name

Agent Task Generator - Optimized Multi-Agent Task Prompt Creation

Description

This skill generates optimized, context-rich task prompts for specialized agents based on the proven template structure from /home/fitna/homelab/sprints/sprint-26/agent-prompts-optimized/. It creates 80-120 line prompts with shared context references, clear objectives, measurable requirements, and copy-pasteable execution instructions. The skill ensures agents receive comprehensive, actionable task definitions that enable autonomous execution.

When to Use This Skill

Trigger Conditions

Use this skill when the user requests ANY of the following:

•"Create an agent prompt for [task]"
•"Generate a task for [AgentType]"
•"Write an optimized prompt for [objective]"
•"Prepare agent instructions for [deliverable]"
•"Create a task delegation prompt"
•"Generate agent-specific task definition"

Context Indicators

•User mentions specific agent types (ContentCreator, Researcher, Analyst, DeploymentOrchestrator, Verifier, ProjectManager)
•User discusses task delegation or multi-agent workflows
•User wants to create reusable task templates
•User references sprint planning or task allocation

Process Steps

Phase 1: Task Analysis (2-3 minutes)

•
Identify Agent Type Available agents:
- •ContentCreatorAgent: Copy, scripts, social media, newsletters
- •ResearcherAgent: Analysis, trends, competitors, communities
- •AnalystAgent: Dashboards, metrics, KPIs, tracking
- •DeploymentOrchestrator: Infrastructure, CI/CD, deployments
- •VerifierAgent: Testing, QA, security, code review
- •ProjectManagerAgent: Planning, coordination, reporting
•
Extract Task Requirements Ask user (if not provided):
- •What is the objective? (1-2 sentences)
- •What are the deliverables? (specific file paths)
- •What is the priority? (HIGH/MEDIUM/LOW/CONTINUOUS)
- •What is the deadline? (date or sprint milestone)
- •What context is needed? (project, sprint, brand, constraints)
•
Define Success Criteria Create 3-5 measurable checkboxes:
- • Deliverable X created at path Y
- • Meets quality standard Z
- • Reviewed and approved
- • Deployed/published (if applicable)

Phase 2: Shared Context Setup (3-5 minutes)

•
Reference Shared Context Files Location: /home/fitna/homelab/sprints/sprint-{NUMBER}/shared-context/

Required Files:
- •@project-context.md - Homelab architecture, goals, constraints
- •@sprint-context.md - Current sprint objectives, timeline, priorities
- •@brand-guidelines.md - Brand voice, tone, visual identity
- •@output-structure.md - Deliverable formatting standards

•

Create Sprint-Specific Context (if not exists)

bash

SPRINT_DIR="/home/fitna/homelab/sprints/sprint-{NUMBER}/shared-context"
mkdir -p "$SPRINT_DIR"

# Copy from previous sprint as template
cp -r /home/fitna/homelab/sprints/sprint-26/shared-context/* "$SPRINT_DIR/"

# Update sprint-context.md with current sprint goals

Phase 3: Prompt Template Generation (5-10 minutes)

•

Create Metadata Section

markdown

---
agent: [AgentType]
model: claude-sonnet-4.5
priority: [HIGH/MEDIUM/LOW/CONTINUOUS]
deadline: [YYYY-MM-DD or "End of Sprint X"]
output_path: /home/fitna/homelab/[category]/sprint-{NUMBER}/
---

•

Write Objective Section

•1-2 sentences max
•Action-oriented (Create, Analyze, Deploy, Verify)
•Specific and measurable

Example:

markdown

## Objective
Create a comprehensive landing page copy that explains Fitnaai's value proposition, targets early adopters, and drives beta signups with clear CTAs and social proof.

•

Define Requirements Section (5-10 items)

•Each requirement is a bullet point
•Must be verifiable (yes/no check)
•Include constraints (word count, format, dependencies)

Example:

markdown

## Requirements
1. Landing page copy (800-1200 words) with hero, features, testimonials, pricing, FAQ sections
2. 3 distinct value propositions targeting fitness coaches, gym owners, and personal trainers
3. Compelling hero headline (8-12 words) and subheadline (15-25 words)
4. 5-7 feature descriptions with icon suggestions (50-80 words each)
5. 3 fictional testimonials with realistic names, titles, and quotes (60-100 words each)
6. Clear CTA buttons (primary: "Start Free Trial", secondary: "Watch Demo")
7. FAQ section answering 8-10 common objections
8. SEO-optimized meta title (50-60 chars) and description (150-160 chars)
9. Tone: Professional, motivating, trustworthy (see @brand-guidelines.md)
10. No marketing jargon, focus on tangible outcomes

•

Add Task-Specific Context (10-20 lines max)

•Relevant background information
•Constraints or dependencies
•Reference materials

Example:

markdown

## Context
Fitnaai is an AI-powered fitness platform for coaches to automate client onboarding, workout planning, and progress tracking. We're launching beta in 4 weeks with 10 invited fitness professionals. Competitors include Trainerize ($$$) and MyPT Hub ($$), but we differentiate with AI-driven personalization at 1/3 the price.

Target audience: Fitness coaches managing 10-50 clients, tech-savvy, struggling with manual admin work, willing to try new tools. Pain points: time spent on repetitive tasks, client retention, workout customization at scale.

Brand voice (from @brand-guidelines.md): We're the smart, efficient partner that frees coaches to focus on what they love—transforming lives. We don't overpromise, we deliver measurable time savings (avg. 10h/week).

References:
- Competitor analysis: /home/fitna/homelab/shared/research/sprint-26/competitor-analysis.md
- User personas: /home/fitna/homelab/shared/docs/user-personas.md
- Product features: /home/fitna/homelab/shared/docs/fitnaai-features.md

•

Define Deliverables Section (exact file paths)

markdown

## Deliverables
Create the following files in `/home/fitna/homelab/shared/content/sprint-26/landing-page/`:

1. `hero-section.md` - Hero headline, subheadline, CTA (200-300 words)
2. `features-section.md` - 5-7 feature blocks with descriptions (400-600 words)
3. `testimonials-section.md` - 3 testimonials with names and titles (200-300 words)
4. `faq-section.md` - 8-10 Q&A pairs (400-600 words)
5. `meta-tags.md` - SEO title, description, social share text (100 words)
6. `full-landing-page.md` - Complete assembled copy (1000-1200 words)

•

Create Success Criteria Section (checkboxes)

markdown

## Success Criteria
- [ ] All 6 deliverable files created with exact word counts
- [ ] Hero headline A/B tested with 3 variants (select best)
- [ ] Features aligned with product roadmap (verified against fitnaai-features.md)
- [ ] Testimonials sound authentic (no generic praise, specific outcomes)
- [ ] FAQs address top 10 objections from competitor reviews
- [ ] SEO meta tags optimized (checked with Yoast/SEMrush)
- [ ] Tone consistency score >90% (checked against brand-guidelines.md)
- [ ] Copy reviewed by ProjectManagerAgent for clarity

•
Add Execution Prompt Section (copy-pasteable)
markdown
```
## Execute This Task

Copy and paste the following into your AI agent interface:
```
You are ContentCreatorAgent. Your task is to create a landing page copy for Fitnaai's beta launch.

Read the following context files first:
- •@project-context.md (homelab architecture and Fitnaai goals)
- •@sprint-context.md (Sprint 26 objectives and timeline)
- •@brand-guidelines.md (brand voice and tone standards)
- •@output-structure.md (deliverable formatting)
[Full task context from above]

Create all 6 deliverable files in /home/fitna/homelab/shared/content/sprint-26/landing-page/.

Ensure every success criterion is met. Mark each checkbox when complete.
code

Phase 4: Optimization and Validation (2-3 minutes)

•
Validate Prompt Length
- •Target: 80-120 lines
- •If <80 lines: Add more context or requirements
- •If >120 lines: Condense or split into multiple tasks
•
Check for Ambiguity Review each section:
- • Objective is specific (no vague goals)
- • Requirements are measurable (can be verified)
- • Deliverables have exact file paths
- • Success criteria are binary (yes/no)
- • No assumptions (all context provided)
•
Test Copy-Pasteability
- •Execution prompt should work standalone
- •All @references are valid file paths
- •No placeholders like [TBD] or [FILL IN]

Rules and Constraints

Hard Rules (Must Follow)

•ALWAYS reference shared context files (@project-context.md, @sprint-context.md, etc.)
•Deliverables MUST have exact file paths (no generic "create a document")
•Success criteria MUST be measurable (checkboxes with verifiable outcomes)
•Prompt length: 80-120 lines (excluding execution block)
•Agent type MUST match task category (don't assign content creation to DeploymentOrchestrator)
•Priority MUST be set (HIGH/MEDIUM/LOW/CONTINUOUS)
•Output path MUST exist or be creatable (verify directory structure)

Soft Rules (Best Practices)

•Limit requirements to 5-10 items (more = split into subtasks)
•Keep context section concise (10-20 lines max)
•Include 3-5 success criteria (not exhaustive)
•Reference existing files for context (avoid copy-pasting large blocks)
•Use concrete examples in requirements (not abstract descriptions)

Quality Gates

Before finalizing prompt:

• Read by someone unfamiliar with the task → is it clear?
• Can success be measured objectively?
• Are all dependencies identified?
• Is the priority justified?
• Is the deadline realistic?

Expected Outputs

Deliverables

•Agent Task Prompt (.md file) - Optimized 80-120 line prompt
•Shared Context Files (if created) - 4 context files for sprint
•Execution Instructions - Copy-pasteable prompt block

File Structure

code

/home/fitna/homelab/sprints/sprint-{NUMBER}/agent-prompts-optimized/
├── ContentCreatorAgent-task-01-landing-page.md
├── ResearcherAgent-task-01-competitor-analysis.md
├── AnalystAgent-task-01-kpi-dashboard.md
├── DeploymentOrchestrator-task-01-cicd-pipeline.md
├── VerifierAgent-task-01-testing-framework.md
└── ProjectManagerAgent-task-01-sprint-planning.md

Success Metrics

•✅ Prompt length: 80-120 lines
•✅ All metadata fields populated
•✅ 5-10 measurable requirements
•✅ Exact deliverable file paths
•✅ 3-5 success criteria (checkboxes)
•✅ Copy-pasteable execution block
•✅ No ambiguous language

Error Handling

Common Issues & Solutions

Issue: User provides vague objective

•Solution: Ask clarifying questions (What exactly? By when? For whom?)
•Example: "Create content" → "Create what type? For which audience? With what goal?"

Issue: Too many requirements (>10)

•Solution: Split into multiple subtasks or prioritize top 5-7
•Action: Suggest breaking down into sprint phases

Issue: Missing shared context files

•Solution: Create from template (/sprints/sprint-26/shared-context/)
•Fallback: Embed context directly in prompt (increases length)

Issue: Deliverable path doesn't exist

•Solution: Create directory structure first or instruct agent to create
•Command: mkdir -p /path/to/deliverables/

Issue: Success criteria not measurable

•Solution: Rephrase as verifiable checkboxes
•Bad: "Make it good" → Good: "Tone consistency score >90%"

Example Usage

User Request

"Create a prompt for ResearcherAgent to analyze competitors for Fitnaai."

Agent Response

markdown

I'll create an optimized agent task prompt for competitor analysis.

**Agent:** ResearcherAgent
**Objective:** Analyze top 5 fitness coaching platforms to identify differentiation opportunities for Fitnaai's beta launch.

**Deliverables:**
- competitor-analysis.md (1500-2000 words)
- competitive-matrix.csv (feature comparison)
- positioning-recommendations.md (500-700 words)

**Output:** `/home/fitna/homelab/sprints/sprint-27/agent-prompts-optimized/ResearcherAgent-task-01-competitor-analysis.md`

✅ Prompt created (95 lines)
✅ 7 measurable requirements
✅ 5 success criteria
✅ Shared context references (@project-context.md, @brand-guidelines.md)
✅ Copy-pasteable execution block included

Ready for ResearcherAgent to execute!

Integration Points

Related Skills

•Agent 1 - SprintManager: Use to allocate generated prompts to sprint board
•Agent 4 - DailyProgressReporter: Track prompt execution progress
•Agent 8 - SOPExecutor: Execute prompts as standard procedures

External Tools

•Shared context files: /home/fitna/homelab/sprints/sprint-{NUMBER}/shared-context/
•Agent prompt templates: /home/fitna/homelab/sprints/sprint-26/agent-prompts/
•Output directories: /home/fitna/homelab/shared/{content,research,dashboards}/

Data Sources

•Agent definitions: /home/fitna/homelab/AGENTS.md
•Sprint plan: /home/fitna/homelab/sprints/sprint-{NUMBER}/SPRINT_PLAN.md
•Roadmap: /home/fitna/homelab/ROADMAP_1M_18MONTHS.md

Version

v1.0 - Initial skill definition based on Sprint 26 agent prompt analysis