CV Knowledge Query

<when_to_activate> Activate when the user:

• Asks "What have I done with [technology/skill]?"
• Wants to find achievements for a specific theme
• Needs metrics for a particular company/period
• Asks "Show me my [type] experience"
• Wants to understand relationships between experiences

Trigger phrases: "find", "search", "what's my experience", "show me", "list achievements" </when_to_activate>

Knowledge Base Schema

Entities (content/knowledge/index.yaml)

entities:
  companies:
    - id: string       # anchorage, microsoft, etc.
      name: string
      period: string
      role: string
      domain: string

  themes:
    - id: string       # institutional-crypto, developer-experience
      label: string
      keywords: [string]

  skills:
    - id: string       # ethereum, api-design
      label: string
      category: string

Relationships

relationships:
  - from: achievement:id
    to: company:id | skill:id | theme:id
    type: achieved_at | demonstrates | belongs_to | contains | generated_from

Achievements (content/knowledge/achievements/*.yaml)

id: string
headline: string
metric:
  value: string
  unit: string
  context: string
situation: string
task: string
action: string
result: string
skills: [string]
themes: [string]
companies: [string]
years: [number]
good_for: [string]

Quick Search (Deterministic Script)

For keyword-based searches, USE THE SCRIPT FIRST — it's faster and consistent:

# Search by keywords
npm run search:evidence -- --terms "crypto,staking,institutional"

# With JSON output for processing
npm run search:evidence -- --terms "crypto,staking" --json

The script returns:

• Matching achievements and stories sorted by relevance
• Relevance scores (strong ≥70%, moderate 40-70%, weak <40%)
• Matched terms, skills, and themes
• Snippets for quick review

Use the script when:

• User asks about specific keywords
• Quick search without deep analysis needed
• Need alignment scoring for job fit

Use manual querying when:

• Deep relationship traversal needed
• Need full STAR narrative (situation, task, action, result)
• Complex cross-referencing across entities

---

Query Patterns

Query by Theme

User: "What's my institutional crypto experience?"

Script-first approach:

npm run search:evidence -- --terms "institutional,crypto,custody"

For deep dive:

1. Find theme institutional-crypto in index
2. Query relationships where to: theme:institutional-crypto
3. Load matching achievements
4. Return summary with metrics

Output Format:

## Institutional Crypto Experience

### Achievements:
1. **ETH Staking Zero Slashing** (Anchorage, 2024-25)
   - Zero slashing events, Galaxy + institutional clients
   - Skills: ethereum, staking, compliance

2. **L2 Protocol Integrations** (Anchorage, 2024-25)
   - 7+ protocols shipped, 40% faster integration
   - Skills: l2s, compliance

Query by Skill

User: "Show me my API design work"

Process:

1. Find skill api-design in index
2. Query achievements where skills includes api-design
3. Return with context and metrics

Query by Company

User: "What did I accomplish at Microsoft?"

Process:

1. Find company microsoft in index
2. Query achievements where companies includes microsoft
3. Include period and role context

Query by Keyword

User: "Find achievements related to revenue growth"

Script-first approach:

npm run search:evidence -- --terms "revenue,growth,ARR"

For manual search:

1. Search theme keywords for "revenue"
2. Search achievement headlines/results for "revenue"
3. Return ranked by relevance

Query for Role Fit

User: "What achievements fit a Platform PM role?"

Script-first approach:

npm run search:evidence -- --terms "platform,infrastructure,api,scale"

For deeper analysis:

1. Parse role keywords: platform, infrastructure, scale
2. Match themes: infrastructure, developer-experience
3. Match good_for arrays in achievements
4. Return ranked list with reasoning

Response Formats

Summary Format (default)

Found 3 achievements matching "[query]":

1. **[Headline]** ([Company], [Year])
   Metric: [value] [unit]
   Themes: [theme1], [theme2]

2. ...

Detail Format (on request)

## [Achievement Headline]

**Context**: [Company] | [Period] | [Role]
**Impact**: [metric.value] [metric.unit]

### Situation
[situation text]

### Task
[task text]

### Action
[action text]

### Result
[result text]

**Best for**: [good_for items]

Metrics Format (for comparisons)

| Achievement | Metric | Value | Company | Year |
|------------|--------|-------|---------|------|
| [headline] | [unit] | [value] | [company] | [year] |

Cross-Reference Queries

Find Related Achievements

User: "What else relates to the ETH staking work?"

Process:

1. Load eth-staking-zero-slashing achievement
2. Find shared themes: institutional-crypto, infrastructure
3. Find shared skills: ethereum, compliance
4. Query other achievements with overlap
5. Rank by relationship strength

Find Gaps

User: "What skills don't have strong achievements?"

Process:

1. List all skills in index
2. Count achievements per skill
3. Flag skills with <2 achievements
4. Suggest areas for content development

Query Shortcuts

User Says	Query
"my blockchain work"	themes: [enterprise-blockchain, institutional-crypto]
"growth stories"	themes: [revenue-growth]
"developer tools"	themes: [developer-experience]
"recent work"	years: [2024, 2025]
"Microsoft days"	companies: [microsoft]
"startup experience"	companies: [ankr, forte, dapper]

<file_locations>

File Locations

Query Target	File Path
Entity definitions	content/knowledge/index.yaml
Achievements	content/knowledge/achievements/*.yaml
Stories	content/knowledge/stories/*.yaml
Metrics	content/knowledge/metrics/*.yaml
</file_locations>

<skill_compositions>

Works Well With

• generate-variant — Query evidence before generating variants
• cv-content-generator — Find achievements for case studies
• generate-story-bank — Search for stories by category </skill_compositions>