/recommend

/recommend Command: Investment Decision & Path Selection

Usage

/recommend

Overview

The /recommend command produces your final investment recommendations and decision paths. It automatically reads your assessment findings and sensitivity analysis from the workspace, and generates two investment pathways: Path A (standard approach) is always available; Path B (alternative approach) is available if you qualified in sensitivity analysis.

What you get: A comprehensive final recommendations document with Path A and Path B options, implementation roadmaps for each, and a complete session audit trail—your final deliverable from the full assessment workflow.

---

Output Folder Paths

All outputs are written to:

Output type	Destination
HTML, PDF reports	$WORKSPACE/assessment/recommendations/reports/
JSON data files	$WORKSPACE/assessment/recommendations/data/

Base paths (used throughout this command):

• Reports: $WORKSPACE/assessment/recommendations/reports/
• Data: $WORKSPACE/assessment/recommendations/data/

---

Pre-Flight Checks

Step 0: Discover workspace root and verify assessment + sensitivity outputs (run first)

Run the following Python script:

import os, glob, json

mounts = glob.glob('/sessions/*/mnt/*/')
workspace = mounts[0].rstrip('/') if mounts else ''
print(f'WORKSPACE={workspace}')

if not workspace:
    print('ERROR: No workspace found. Please select a folder in Cowork mode.')
else:
    assess_data = os.path.join(workspace, 'assessment', 'assessment', 'data')
    sens_data = os.path.join(workspace, 'assessment', 'sensitivity', 'data')
    
    if not os.path.isdir(assess_data):
        print('ERROR: assessment/data not found. Run /assess first.')
    elif not os.path.isdir(sens_data):
        print('ERROR: sensitivity/data not found. Run /sensitivity first.')
    else:
        cp = os.path.join(workspace, 'assessment', 'pre-assessment', 'data', 'context-profile.json')
        if os.path.exists(cp):
            with open(cp) as fh:
                data = json.load(fh)
            print(f'Company: {data.get("company_name", "unknown")}')
        print(f'assessment/data files: {os.listdir(assess_data)}')
        print(f'sensitivity/data files: {os.listdir(sens_data)}')

Use the WORKSPACE value above for all file paths. If WORKSPACE is empty or prerequisite folders are missing, stop and follow the error instructions.

---

Step 1: Read all required files into context (mandatory — do this before anything else)

Use the Read tool to open and read every file listed below. Do not invoke the recommendations-agent until all files are fully loaded into context.

Read these files in order:

1. $WORKSPACE/assessment/pre-assessment/data/context-profile.json
2. $WORKSPACE/assessment/pre-assessment/data/assessor-profile.json
3. $WORKSPACE/assessment/pre-assessment/data/framework.json
4. $WORKSPACE/assessment/pre-assessment/data/gap-register.json
5. $WORKSPACE/assessment/assessment/data/updated-go-nogo-determination.json
6. $WORKSPACE/assessment/assessment/data/integrated-findings-register.json
7. All $WORKSPACE/assessment/assessment/data/domain-findings-*.json files — use the Read tool on each one found
8. $WORKSPACE/assessment/sensitivity/data/sensitivity-analysis.json

Also read any CP review files in $WORKSPACE/assessment/assessment/reports/ and $WORKSPACE/assessment/sensitivity/reports/ for full audit trail context.

Confirm in chat: "Loaded all data for [Company Name] — Path B eligible: [yes/no]" before proceeding.

---

Step 1: Recommendations Generation

Agent: recommendations-agent

• Input (all read automatically from workspace):
  • $WORKSPACE/assessment/assessment/data/integrated-findings-register.json
  • $WORKSPACE/assessment/assessment/data/updated-go-nogo-determination.json
  • $WORKSPACE/assessment/assessment/data/domain-findings-[domain_id].json (all domain files)
  • $WORKSPACE/assessment/sensitivity/data/sensitivity-analysis.json
  • $WORKSPACE/assessment/pre-assessment/data/framework.json
  • $WORKSPACE/assessment/pre-assessment/data/context-profile.json
  • $WORKSPACE/assessment/pre-assessment/data/assessor-profile.json
  • $WORKSPACE/assessment/pre-assessment/data/gap-register.json
• Reads Path B eligibility from sensitivity-analysis.json
• Constructs recommendations:

Path A (Always Available):

• Generates standard investment recommendation
• Recommendation level: GO / CONDITIONAL GO / PASS
• Rationale: rooted in assessment findings and final determination
• Investment approach:
  • Deal structure (equity size, valuation frame, governance)
  • Key conditions precedent (if CONDITIONAL GO)
  • Monitoring requirements (key metrics, board observation rights, etc.)
  • Value-add strategy (introductions, recruitment support, operational guidance, etc.)
• Implementation roadmap:
  • Week 1–2: due diligence completion and condition verification
  • Week 3–4: term sheet negotiation and signing
  • Week 5–6: closing and deployment
  • Post-close: monitoring, support, governance participation
• Success criteria: metrics that indicate investment is tracking
• Risk mitigation: top 3 risks identified in assessment; how they'll be monitored and managed

Path B (If Eligible):

• Generates alternative investment recommendation
• Reflects structured approach or alternative strategy from sensitivity analysis
• May include:
  • Tranched investment structure (initial tranche + additional tranches tied to milestones)
  • Earnout conditions (additional investment or enhanced terms based on performance)
  • Partnership opportunities (non-dilutive revenue sources, strategic relationships)
  • Extended timeline (phased entry with option to increase)
  • Governance provisions (advisory board seats, operational metrics reporting)
• Path B is designed to address conditions or risks that were identified in assessment
• Implementation roadmap specific to Path B structure
• Success metrics and gating criteria for Path B progression

recommendations-agent:

• Determines Path A recommendation level based on final assessment determination
• Determines Path B structure and recommendation level (if Path B eligible)
• Cross-checks Path B rationale against assessment and sensitivity findings
• Produces $WORKSPACE/assessment/recommendations/data/recommendations.json:
  • path_a_recommendation (GO / CONDITIONAL GO / PASS)
  • path_a_rationale
  • path_a_deal_structure
  • path_a_conditions (if CONDITIONAL GO)
  • path_a_monitoring_framework
  • path_a_implementation_roadmap
  • path_a_value_add_strategy
  • path_b_eligible (true / false)
  • path_b_recommendation (if eligible) (GO / CONDITIONAL GO)
  • path_b_rationale
  • path_b_structure
  • path_b_gating_criteria
  • path_b_implementation_roadmap
  • comparative_analysis (if both paths available): pros/cons of each approach

---

Step 2: Output Generation

Agent: recommendations-agent (continued, output generation mode)

• Input (all read automatically from workspace):
  • $WORKSPACE/assessment/recommendations/data/recommendations.json
  • $WORKSPACE/assessment/assessment/data/integrated-findings-register.json
  • $WORKSPACE/assessment/assessment/data/updated-go-nogo-determination.json
  • $WORKSPACE/assessment/sensitivity/data/sensitivity-analysis.json
  • $WORKSPACE/assessment/pre-assessment/data/context-profile.json
  • $WORKSPACE/assessment/pre-assessment/data/assessor-profile.json
  • Session audit trail (CP1–CP5 confirmations and all adjustments)
• Mandatory: Load the design-system skill and the html-dashboard skill before generating. Apply the centralized design system's tokens and meet the Quality Contract. Adapt tone to assessor type. This is the final deliverable — quality standards are at their highest here.
• Content freedom: The agent determines optimal structure, sections, and narrative for this specific case — which paths to present, how to frame the roadmap, what deal terms to suggest, how to tell the investment story.
• Generates 3 deliverable outputs (all saved to $WORKSPACE/assessment/recommendations/reports/):

1. [CompanyName]Recommendations[YYYY-MM-DD].html

   • Self-contained interactive HTML report using the html-dashboard skill's component library and chart patterns
   • Content adaptive to determination outcome, assessor type, and specific case findings — path recommendations, deal structures, roadmaps, success criteria as appropriate
   • Must meet the design system's quality contract at the highest level — this report goes to investment committees, credit committees, boards, and LPs

2. [CompanyName]Recommendations[YYYY-MM-DD].pdf

   • Print-optimized HTML variant; format adapts to assessor type per html-dashboard skill
   • FINAL DELIVERABLE — includes complete session audit trail in appendix
   • Audit trail documents:
     • CP1 (Context & Assessor Profile) confirmation
     • CP2 (Framework) confirmation and any adjustments
     • CP3 (Scored Findings) confirmation and any notes
     • CP4 (Assessment Scope) confirmation and any escalations
     • CP5 (Reconciled Findings) confirmation and any overrides
     • All research sources and confidence scores
     • All scoring logic and determination pathways
     • All sensitivity methodology and Path B eligibility logic
   • Ready for regulatory filing, LP reporting, or deal documentation

3. recommendations.json → saved to $WORKSPACE/assessment/recommendations/data/

   • Machine-readable recommendations export
   • Contains all structured recommendation data, paths, and rationale
   • Useful for downstream workflow integration

---

Workflow Completion

Output locations:

assessment/recommendations/
├── reports/
│   ├── [CompanyName]_Recommendations_[YYYY-MM-DD].html
│   └── [CompanyName]_Recommendations_[YYYY-MM-DD].pdf  ← FINAL DELIVERABLE
└── data/
    └── recommendations.json

Summary of All Deliverables:

You now have a complete assessment file set across all 4 phases:

Pre-Assessment Phase (5 files):

1. [CompanyName]PreAssessment[YYYY-MM-DD].html
2. [CompanyName]PreAssessment[YYYY-MM-DD].pdf
3. [CompanyName]PreAssessment[YYYY-MM-DD].md
4. [CompanyName]ScoredRegister[YYYY-MM-DD].json
5. [CompanyName]ResearchProvenance[YYYY-MM-DD].json

Assessment Phase (3 files): 6. [CompanyName]Assessment[YYYY-MM-DD].html 7. [CompanyName]Assessment[YYYY-MM-DD].pdf 8. [CompanyName]Assessment[YYYY-MM-DD].md

Sensitivity Phase (3 files): 9. [CompanyName]Sensitivity[YYYY-MM-DD].html 10. [CompanyName]Sensitivity[YYYY-MM-DD].pdf 11. [CompanyName]Sensitivity[YYYY-MM-DD].md

Recommendations Phase (3 files): 12. [CompanyName]Recommendations[YYYY-MM-DD].html 13. [CompanyName]Recommendations[YYYY-MM-DD].pdf ← FINAL DELIVERABLE with complete audit trail 14. recommendations.json

Total: 14 files across all phases

---

Congratulations! You have completed the full startup assessment workflow. Your Recommendations PDF is your final deliverable—it contains your investment decision, implementation roadmap, and complete session audit trail documenting every confirmation point and adjustment made throughout the process.

---

Next Steps:

• If you have an updated business case or submission: You may run /pre-assess again at any time with the revised document. This will create a new assessment session with fresh findings, allowing you to track changes over time.

• If you want to explore alternative scenarios: You can re-run /sensitivity with different methodology options, or adjust assessment scope and re-run /assess to test alternative assessment hypotheses.

• For external sharing: All PDF outputs are ready for investor review, board presentation, or partnership discussions. The Recommendations PDF includes complete audit trail for transparency and compliance.