Biographical Research

Research individuals: $ARGUMENTS

System Prompt

⚠️ EXECUTION DIRECTIVE: When the user invokes this command, you MUST:

1. IMMEDIATELY execute - no questions, no explanations first
2. ALWAYS show full output from script/tool execution
3. ALWAYS provide summary after execution completes

DO NOT:

• Say "I don't need to take action" - you ALWAYS execute when invoked
• Ask for confirmation unless requires_confirmation: true in frontmatter
• Skip execution even if it seems redundant - run it anyway

The user invoking the command IS the confirmation.

---

Arguments

$ARGUMENTS - Research Configuration (required)

Specify individuals to research:

• Names + context: "--names 'Name1, Name2' --context 'Company X'"
• With focus: "--names 'Name1' --context 'Startup' --focus 'education, publications'"
• Thoroughness: "--names 'Name1' --context 'Investor' --thoroughness comprehensive"
• Custom output: "--names 'Name1' --context 'CEO' --output profiles.md --format json"

Default Behavior

Requires --names and --context arguments. No automatic defaults due to sensitive nature of biographical research.

Description

Execute systematic biographical research on individuals (founders, executives, investors, researchers) using structured web search methodology with multi-source validation.

Usage

/research-bio --names "Name1, Name2, Name3" --context "Company X" [options]

Parameters

Required

• --names - Comma-separated list of names to research
• --context - Company, project, or organizational context

Optional

• --focus - Areas to emphasize (education, publications, investments, etc.)
• --output - Output file path (default: auto-generated markdown)
• --format - Output format: markdown (default), json, csv
• --thoroughness - Level: basic (15-20 min), standard (30-45 min), comprehensive (45-60 min)
• --validate - Validation level: standard (2+ sources), strict (3+ sources), relaxed (1+ source)
• --confidence-threshold - Minimum confidence to include profile: high (90%+), medium (70%+), low (50%+)

Examples

Example 1: Research Startup Founding Team

/research-bio \
  --names "Tamay Besiroglu, Matthew Barnett, Ege Erdil" \
  --context "Mechanize.work" \
  --focus "education, epoch-ai-background, publications" \
  --thoroughness comprehensive \
  --output docs/mechanize-founders.md

Output: Comprehensive profiles for all three founders with emphasis on educational credentials, Epoch AI work, and research publications.

Example 2: Research Investor Syndicate

/research-bio \
  --names "Nat Friedman, Daniel Gross, Patrick Collison" \
  --context "Mechanize.work investors" \
  --focus "portfolio, investment-thesis, partnerships" \
  --thoroughness standard \
  --output docs/mechanize-investors.md

Output: Investor profiles highlighting portfolio companies, investment focus, and co-investment patterns.

Example 3: Quick Research for Meeting Prep

/research-bio \
  --names "Jane Doe" \
  --context "ABC Company CEO" \
  --thoroughness basic \
  --focus "education, current-role"

Output: Basic profile (15-20 min) with education and current role for quick meeting preparation.

Example 4: Research Academic Collaborators

/research-bio \
  --names "Prof. Smith, Dr. Jones, Prof. Lee" \
  --context "AI Research Collaboration" \
  --focus "publications, research-areas, h-index" \
  --thoroughness comprehensive \
  --validate strict

Output: Academic profiles with emphasis on research output, requiring 3+ sources per claim.

Example 5: Export to JSON for Integration

/research-bio \
  --names "Person1, Person2, Person3" \
  --context "Company X Team" \
  --format json \
  --output team_profiles.json

Output: JSON-formatted profiles for programmatic integration.

Workflow

When you invoke /research-bio, the following workflow executes:

Step 1: Parse Input

Parsing command parameters:
- Names: [List]
- Context: [Company/Project]
- Focus areas: [Specified or default]
- Thoroughness: [basic/standard/comprehensive]
- Validation level: [relaxed/standard/strict]

Step 2: Invoke biographical-researcher Agent

Task(subagent_type="biographical-researcher",
     description="Research biographical profiles",
     prompt=f"""
     Research the following individuals:

     Names: {names}
     Context: {context}
     Focus Areas: {focus_areas}
     Thoroughness: {thoroughness}

     For each individual:
     1. Confirm identity and disambiguate
     2. Research: {phase_list}
     3. Validate with {validation_level} (sources per claim)
     4. Compile structured profile with citations
     5. Note confidence level and information gaps

     Output Format: {format}
     Output File: {output_file}
     """)

Step 3: Execute Research Phases

Phase 1: Initial Identification (5-10 min)

• Confirm identity
• Establish baseline (affiliation, location, role)

Phase 2: Educational Background (10-15 min)

• Degrees and institutions
• Fields of study
• Academic achievements

Phase 3: Professional Experience (15-20 min)

• Current and previous roles
• Career trajectory
• Notable achievements

Phase 4: Research & Publications (15-20 min, if applicable)

• Google Scholar profile
• Notable papers
• Citation metrics

Phase 5: Investment History (10-15 min, if applicable)

• Portfolio companies
• Investment focus
• Partnership patterns

Phase 6: Public Presence (10-15 min)

• Social media, websites
• Interviews, podcasts
• Public statements

Step 4: Validation & Quality Check

• Verify 2+ (standard) or 3+ (strict) sources per claim
• Confirm identity disambiguation
• Check date accuracy
• Assess information freshness

Step 5: Compile Output

Generate structured profiles in specified format with:

• All sections completed
• Citations for all claims
• Confidence levels noted
• Information gaps documented

Step 6: Generate Summary Report

Create research methodology report including:

• Search strategy used
• Disambiguation challenges
• Information gaps
• Confidence assessment
• Recommendations for follow-up

Output Structure

Markdown Format (Default)

# [Context] - Biographical Profiles

**Research Date:** [Date]
**Researcher:** biographical-researcher agent
**Validation:** [Level]

---

## [Individual 1 Name]

[Complete profile per agent spec]

---

## [Individual 2 Name]

[Complete profile per agent spec]

---

## Research Methodology

### Search Strategy
[Summary of approach]

### Challenges Encountered
[Disambiguation issues, gaps, etc.]

### Information Gaps
[What couldn't be verified]

### Confidence Levels
- [Name 1]: HIGH (95%+ confident in identity and claims)
- [Name 2]: MEDIUM (70-89% confident)
- [Name 3]: LOW (50-69% confident)

---

## Sources Bibliography

1. [Source 1 - URL]
2. [Source 2 - URL]
...

JSON Format

{
  "research_date": "2025-12-02",
  "context": "Company X",
  "validation_level": "standard",
  "profiles": [
    {
      "name": "Person 1",
      "role": "Co-Founder",
      "affiliation": "Company X",
      "education": [...],
      "professional_background": [...],
      "research": {...},
      "public_presence": {...},
      "company_connection": {...},
      "sources": [...],
      "confidence": "HIGH",
      "completeness": 0.85
    }
  ],
  "methodology": {
    "search_strategy": "...",
    "challenges": [...],
    "gaps": [...]
  },
  "sources": [...]
}

Focus Area Presets

For Founders/Executives

--focus "education, professional-background, company-history, public-statements"

For Investors

--focus "portfolio, investment-thesis, partnerships, exits"

For Researchers

--focus "education, publications, research-areas, h-index, collaborators"

For All Roles

--focus "education, professional-background, public-presence, company-connection"

Thoroughness Levels

Basic (15-20 min per person)

• Identity confirmation
• Current role
• Education (degrees only)
• LinkedIn profile
• 3-5 sources minimum

Use when: Quick meeting prep, basic due diligence

Standard (30-45 min per person)

• Complete educational background
• Professional history (all roles)
• Public presence (social media, etc.)
• 5-7 sources minimum

Use when: Standard research, investor profiles, general due diligence

Comprehensive (45-60 min per person)

• All standard items
• Research publications (if applicable)
• Investment portfolio (if applicable)
• Media appearances and interviews
• Co-authorship/collaboration networks
• 7-10 sources minimum

Use when: Founder research, academic collaborators, competitive intelligence

Validation Levels

Relaxed (1+ source)

• Single authoritative source sufficient
• Faster research (50-60% time savings)
• Accept some uncertainty
• Mark all claims with confidence levels

Use when: Time-constrained, low-stakes research, preliminary investigation

Standard (2+ sources)

• Default validation level
• Balance speed and accuracy
• 2 independent sources per major claim
• Recommended for most use cases

Use when: General research, standard due diligence, investor profiles

Strict (3+ sources)

• 3+ independent sources per claim
• Highest accuracy
• Slower research (+30-40% time)
• Zero fabrication tolerance

Use when: High-stakes research, competitive intelligence, legal/compliance needs

Confidence Thresholds

High (90%+)

• Only include profiles with 90%+ confidence in identity
• Strict disambiguation required
• May exclude ambiguous individuals

Use when: Legal/compliance, high-stakes decisions, public documentation

Medium (70%+)

• Default threshold
• Include profiles with reasonable confidence
• Note ambiguities clearly

Use when: Standard research, internal documentation, preliminary analysis

Low (50%+)

• Include profiles with modest confidence
• Clearly mark uncertainty
• Flag for manual review

Use when: Exploratory research, preliminary investigation, time-constrained scenarios

Error Handling

Scenario: Name Disambiguation Failure

Message:

⚠️ Unable to confidently identify: [Name]
Reason: Multiple individuals with same name found
Suggestion: Add more context (company, location, specialty)

Retry with:
/research-bio --names "[Name]" --context "[More specific context]" --focus "[Distinctive attribute]"

Scenario: Insufficient Information

Message:

⚠️ Limited information available: [Name]
Sources found: [Count] (below threshold)
Confidence: LOW

Profile generated with available data.
Marked for manual follow-up.

Scenario: Conflicting Information

Message:

⚠️ Conflicting information found: [Name]
Conflict: [Description]
Sources: [Source A] vs [Source B]

Profile includes both versions with dates and source citations.
Marked for manual verification.

Integration

With Other Commands

# Combine with competitive analysis
/research-bio --names "Competitor CEOs" --context "Market X" --output team.md
/multi-agent-research --focus "Market X competitors" --use-profiles team.md

# Integrate with business intelligence
/research-bio --names "Investor list" --context "Funding round" --output investors.md
# Use investors.md in deal evaluation

With CODITECT Workflows

# Step 1: Research team
/research-bio --names "Target team" --context "Acquisition target"

# Step 2: Competitive analysis
/analyze --use-profiles docs/target-team-profiles.md

# Step 3: Strategic planning
/strategy --incorporate-team-analysis docs/target-team-profiles.md

Performance

Expected Time

• Basic: 15-20 min per person
• Standard: 30-45 min per person
• Comprehensive: 45-60 min per person

Parallelization

Command supports parallel research via multi-agent orchestration:

• 1-3 individuals: Sequential research
• 4-6 individuals: 2 parallel agents (50% faster)
• 7+ individuals: 3 parallel agents (65% faster)

Quality Metrics

• Source count: 5-7 average
• Completeness: 80%+
• Accuracy: 95%+ (0% fabrication)
• Freshness: <12 months for current roles

Best Practices

DO:

✅ Provide specific context (company, role, location) ✅ Use appropriate thoroughness for use case ✅ Specify focus areas to guide research ✅ Review output for disambiguation issues ✅ Follow up on low-confidence profiles

DON'T:

❌ Use generic context ("Investor" → "AI Investor in SF") ❌ Request comprehensive for all use cases (time-consuming) ❌ Ignore confidence levels in output ❌ Skip manual review of ambiguous profiles ❌ Rely on single-source profiles for critical decisions

Troubleshooting

Issue: "No results found for [Name]"

Solutions:

1. Check spelling and name variations
2. Add more context (company, location)
3. Try alternative names (nicknames, full names)
4. Reduce confidence threshold: --confidence-threshold low

Issue: "Multiple individuals found, unable to disambiguate"

Solutions:

1. Add more specific context
2. Include company name: --context "[Name] at [Company]"
3. Add location: --context "[Name] in [City]"
4. Include specialty: --focus "[Specialty]"

Issue: "Research taking too long"

Solutions:

1. Reduce thoroughness: --thoroughness basic
2. Limit focus areas: --focus "education, current-role"
3. Use relaxed validation: --validate relaxed
4. Research fewer individuals per command

Issue: "Low completeness score (< 60%)"

Solutions:

1. Person may have limited public presence (legitimate)
2. Try alternative sources (GitHub, Medium, personal blog)
3. Search for co-authors, collaborators, or colleagues
4. Accept partial profile and note gaps

Privacy & Ethics

Compliance

• Public information only - No private/confidential data sought
• Professional focus - Focus on professional/public persona
• Source attribution - All sources cited for transparency
• Accuracy priority - No fabrication or speculation

GDPR Considerations

• Professional data (exempted under legitimate interest)
• No sensitive personal data processing
• Mechanism for corrections/removal if requested
• Legal basis documented

Version History

• v1.0 (Dec 2025): Initial command definition

Related Components

• Agent: agents/biographical-researcher.md
• Skill: skills/biographical-research/SKILL.md
• Script: scripts/biographical_researcher.py
• Templates: prompts/biographical-research/

Success Output

When biographical research completes:

✅ COMMAND COMPLETE: /research-bio
Names: N researched
Context: <context>
Thoroughness: <basic|standard|comprehensive>
Validation: <relaxed|standard|strict>
Output: <path-to-profiles>
Confidence: Average X%

Completion Checklist

Before marking complete:

• Identity confirmed for each person
• Sources validated (2+ for standard)
• Profile sections completed
• Citations included
• Confidence levels noted

Failure Indicators

This command has FAILED if:

• ❌ Identity disambiguation failed
• ❌ Below confidence threshold
• ❌ Insufficient sources
• ❌ No profile generated

When NOT to Use

Do NOT use when:

• Private information sought
• Non-professional context
• No public presence exists

Anti-Patterns (Avoid)

Anti-Pattern	Problem	Solution
Generic context	Disambiguation fails	Be specific
Single source reliance	Low accuracy	Use standard validation
Skip confidence check	False information	Review all levels

Principles

This command embodies:

• #9 Based on Facts - Multi-source validation
• #5 No Assumptions - Verify identity first

Full Standard: CODITECT-STANDARD-AUTOMATION.md

---

Command Status: Production-ready Documentation: Complete Last Updated: December 2, 2025