Shadow Scout

You should:

1. Use the Shadow-Scout Python package to run the assessment
2. Present results in a clear, actionable format
3. Highlight key findings and recommendation
4. Provide next steps based on recommendation

Running an Assessment

Use the CLI:

cd /home/user/shadow-scout
python cli.py assess <company_url> --name "Company Name"

Or use the Python API directly:

from shadow_scout.agent import ShadowScoutAgent

agent = ShadowScoutAgent()
result = agent.assess_company(
    company_url="https://company.com",
    company_name="Acme Corp"
)

Interpreting Results

Recommendations

PURSUE = High-value target

• Business qualified ✓
• NICE passed (all 4 criteria present) OR strong SPICE: Everything Nice
• Q2-High/Critical or Q3 quadrant
• Reach out immediately with personalized approach

EXPLORE = Needs more investigation

• Business qualified ✓
• NICE: 1-2 missing OR SPICE: Everything Nice/Else
• Q2 (any pain) or unclear quadrant
• Gather more intelligence before outreach

PASS = Not a fit

• Business NOT qualified OR
• Q1 (Consensus-Locked) OR
• SPICE: Potential Ice OR
• 3-4 NICE criteria missing with red flags

Key Metrics

• NICE Score: "All 4 Present (NICE)" / "1-2 Missing" / "3-4 Missing (Not NICE)"
• SPICE Status: "Everything Nice" / "Everything Else" / "Potential Ice"
• Reality-Map Quadrant: Q1 (not ready) / Q2 (high-value) / Q3 (sophisticated) / Q4 (partner)
• Pain Level: Low / Medium / High / Critical
• Narrative Capacity: Low / Medium / High

Output Format

After running assessment, present:

1. Executive Summary (2-3 paragraphs)

   • Recommendation and confidence level
   • Key findings
   • Most important signals

2. Assessment Details (structured)

   • Business qualification results
   • NICE criteria breakdown
   • Reality-Map classification
   • Supporting evidence

3. Next Steps (actionable)

   • For PURSUE: Outreach strategy and email draft
   • For EXPLORE: Intelligence gaps and research plan
   • For PASS: Rationale and what would need to change

4. Outputs Created

   • Link to detailed markdown report
   • Pipedrive deal URL (if synced)

Configuration

Ensure .env is configured with:

• ANTHROPIC_API_KEY
• PIPEDRIVE_API_KEY
• PIPEDRIVE_DOMAIN

Run python cli.py setup for interactive configuration.

Examples

Example 1: Simple Assessment

User: "Run Shadow-Scout on https://acmecorp.com"

You should:

1. Run: python cli.py assess https://acmecorp.com
2. Review output
3. Present summary with recommendation
4. Show report path and Pipedrive link

Example 2: Assessment with Context

User: "Assess https://techstartup.io - they just raised Series B and the CEO posted about culture challenges"

You should:

1. Run: python cli.py assess https://techstartup.io --name "TechStartup" --context "Just raised Series B, CEO posting about culture challenges"
2. Note the additional context in assessment
3. Present findings highlighting the distress signals

Example 3: Batch Assessment

User: "Assess all companies in targets.csv"

You should:

1. Verify CSV format (url, name, context columns)
2. Run: python cli.py batch targets.csv
3. Present comparative summary
4. Highlight PURSUE recommendations for immediate action

Important Notes

• Public data only: Shadow-Scout uses only publicly available information
• Prompt caching: First assessment slower, subsequent ones faster (90% cost reduction)
• Dry run mode: Use --dry-run flag to preview without writing to Pipedrive
• Reports saved: Markdown reports saved to ./reports/ directory
• Pipedrive sync: Automatic unless --no-pipedrive flag used

Framework Reference

NICE Criteria

• N - Narratives Align: Organizational self-awareness, introspection, acknowledges gaps
• I - Integrity Present: Actions match values, consistency, follow-through
• C - Collaboration Signals: Partnership orientation, open to expertise, growth mindset
• E - Equity Demonstrated: Diverse leadership, substantive inclusion, systemic thinking

Reality-Map Quadrants

• Q1 - Consensus-Locked: Rigid narratives, no complexity awareness → NOT VIABLE
• Q2 - Reality-Friction: Something breaking, searching for language → PRIME TARGET
• Q3 - Narrative-Aware: Meta-awareness, sophisticated culture work → QUALIFIED
• Q4 - Post-Integration: Teaching this work themselves → PARTNERSHIP

SPICE Categories (for companies that don't pass NICE)

• Everything Nice: Positive signals, cautiously optimistic, may still pursue
• Everything Else: Neutral unknown, insufficient data, observe
• Potential Ice: Red flags, concerning patterns, proceed with extreme caution

Troubleshooting

If assessment fails:

1. Check API keys in .env
2. Verify internet connectivity
3. Run python cli.py test to diagnose issues
4. Check company URL is accessible
5. Review error messages for specific issues

For questions about the framework or interpretation, refer to:

• config/nice_framework.md - NICE assessment details
• config/reality_map.md - Reality-Map classification guide
• config/azure_ascent_profile.md - Ideal client profile

Remember: Shadow-Scout is reconnaissance, not decision-making. Provide intelligence, flag limitations, and empower the user's judgment.