Trigger phrases: "create case study", "write blog post", "new content", "generate [content type]"

Knowledge Base Structure

content/knowledge/
├── index.yaml           # Entities and relationships graph
├── achievements/        # Atomic accomplishments (STAR format)
├── stories/            # Extended narratives
├── metrics/            # Quantified results
└── raw/                # Unstructured source material

Content Generation Workflow

Step 1: Understand the Request

Determine output type:

• case-study → Full markdown with frontmatter for content/case-studies/
• blog-post → Markdown with frontmatter for content/blog/
• variant → YAML overrides for content/variants/
• experience-update → YAML additions for content/experience/

Step 2: Query Knowledge Base

Use deterministic scripts first — faster and more consistent:

# Search by topic
npm run search:evidence -- --terms "revenue,growth,api"

# For variant generation with JD analysis
npm run analyze:jd -- --file source-data/jd-{company}.txt --save
npm run search:evidence -- --jd-analysis capstone/develop/jd-analysis/{slug}.yaml

For deeper exploration:

1. Read content/knowledge/index.yaml for entity definitions and relationships
2. Find relevant achievements in content/knowledge/achievements/
3. Find related stories in content/knowledge/stories/
4. Cross-reference themes and skills for comprehensive context

Step 3: Generate Content

For Case Studies

Use this structure:

---
id: [next number]
slug: [kebab-case-slug]
title: [Title]
company: [Company]
year: [Year]
tags: [relevant tags]
duration: [duration]
role: [role]

hook:
  headline: [3-second grab]
  impactMetric:
    value: "[X]"
    label: [metric type]
  subMetrics:
    - value: "[Y]"
      label: [secondary metric]
  thumbnail: null

cta:
  headline: [Call to action question]
  subtext: [Supporting text]
  action: calendly
  linkText: Let's talk →
---

[Opening hook - why this matters, stakes involved]

## The Challenge
[Problem statement with constraints]

## The Approach
[Hypothesis and alternatives considered table]

## Key Decision
[Critical decision point with trade-offs]

## Execution
[Phases with specific actions]

## Results
[Quantified outcomes]

## What I Learned
[Reflections - what worked, what didn't, key quote]

For Blog Posts

---
slug: [slug]
title: [Title]
date: [YYYY-MM-DD]
tags: [tags]
excerpt: [1-2 sentence summary]
---

[Content following narrative structure from stories]

For Variants

metadata:
  company: "[Company]"
  role: "[Role]"
  slug: "[company-role]"
  generatedAt: "[ISO timestamp]"
  jobDescription: "[JD summary]"

overrides:
  hero:
    status: "[Customized status]"
    subheadline: "[Tailored pitch]"
  about:
    tagline: "[Role-specific tagline]"
    bio: [Customized paragraphs]
    stats: [Relevant stats]

relevance:
  caseStudies:
    - slug: "[most relevant]"
      relevanceScore: 0.95
      reasoning: "[Why this matters for role]"

Step 4: Validate Output

• Check all required frontmatter fields
• Ensure metrics are quantified
• Verify skills/themes match knowledge base
• Confirm narrative follows STAR format

Examples

Example 1: Generate Case Study

User: "Create a case study about the Ankr revenue growth"

Action:

1. Read content/knowledge/achievements/ankr-15x-revenue.yaml
2. Read content/knowledge/index.yaml for Ankr relationships
3. Generate full case study markdown
4. Output to content/case-studies/ with proper frontmatter

Example 2: Generate Variant

User: "Create a variant for a Technical PM role at Stripe"

⚠️ For full variant workflow, use the generate-variant skill instead.

Quick variant with scripts:

# 1. Analyze the JD
npm run analyze:jd -- --file source-data/jd-stripe.txt --save

# 2. Search for matching evidence
npm run search:evidence -- --jd-analysis capstone/develop/jd-analysis/stripe.yaml

# 3. Check bullet coverage
npm run check:coverage

Action (manual):

1. Query achievements with themes: [infrastructure, revenue-growth]
2. Find skills: [api-design, compliance]
3. Customize hero/about with payments/fintech angle
4. Score case studies by relevance to Stripe's domain
5. Output YAML + JSON variant files

Output Locations

Quality Checklist

Before outputting content:

• All metrics are specific and quantified
• STAR format is complete (Situation, Task, Action, Result)
• Key quote/insight is memorable
• Tags align with knowledge base themes
• Frontmatter validates against schema