Use this skill when evaluating research projects, POCs, experiments, or prototypes for production readiness. Provides the "What Did We Prove?" framework, production readiness checklist, and gap analysis matrix for systematic research evaluation.
Systematically evaluate research artifacts to determine production readiness.
Before planning production work, separate fact from assumption:
Claims with concrete evidence:
Claims we believe but haven't verified:
Things we don't know:
Score each criterion 1-10:
| Criterion | Weight | Questions to Ask |
|---|---|---|
| Core hypothesis validated | 20% | Did the experiment prove what we set out to prove? |
| Performance benchmarks | 15% | Do we have latency, throughput, resource usage data? |
| Security considerations | 15% | Have we identified and addressed security risks? |
| Scalability tested | 10% | Will it work at 10x, 100x current scale? |
| Error handling | 10% | What happens when things fail? |
| Observability | 10% | Can we monitor and debug in production? |
| Documentation | 10% | Could someone else take this over? |
| Knowledge transfer | 10% | Is knowledge spread across the team? |
Scoring Guide:
Use this template to map what's proven vs unknown:
| Domain | Proven | Assumed | Unknown |
|---|---|---|---|
| Functionality | Feature X works | Feature Y will work similarly | Edge case behavior |
| Performance | 100ms p50 latency | Will scale linearly | Performance under load |
| Security | Auth works | No vulnerabilities | Penetration test results |
| Scalability | Works for 100 users | Will work for 10K | Actual breaking point |
| Operations | Can deploy manually | Automation will work | Rollback procedure |
| Integration | API contract defined | Will integrate smoothly | Error handling at boundaries |
Watch for these patterns that indicate low readiness:
# Research Assessment: [Project Name]
## Executive Summary
[2-3 sentences: what this is, key finding, readiness score]
## What Did We Prove?
### Validated Hypotheses
- **[Hypothesis]**: [Evidence] (Confidence: High/Medium/Low)
### Key Findings
1. [Finding with evidence]
## Production Readiness: X/10
| Criterion | Score | Notes |
|-----------|-------|-------|
| Core hypothesis | X | |
| Performance | X | |
| Security | X | |
| Scalability | X | |
| Error handling | X | |
| Observability | X | |
| Documentation | X | |
| Knowledge transfer | X | |
## Gap Analysis
| Domain | Proven | Assumed | Unknown |
|--------|--------|---------|---------|
| ... | ... | ... | ... |
## Risks & Technical Debt
- [Item with severity]
## Recommended Next Steps
1. [Action]
Before investing in production readiness, projects must pass increasingly rigorous gates. Apply the "80% of experiments should fail" philosophy:
| Gate | Kill Rate | Key Question |
|---|---|---|
| Idea → Internal Build | ~40% | Is the problem real and worth solving? |
| Internal → Private Preview | ~30% | Have we built something worth testing externally? |
| Private → Public Preview | ~30% | Are external users getting value? |
| Public → GA | ~10% | Are we ready to bet the company reputation? |
See references/graduation-criteria.md for detailed gate requirements.
See references/ for:
readiness-criteria.md - Detailed scoring rubricscommon-gaps.md - Typical gaps by project typegraduation-criteria.md - Gate criteria and kill decisions (GitHub Next patterns)rfc-templates.md - Google, Uber, Sourcegraph RFC formatsSee examples/ for:
poc-assessment-example.md - Vector Search POC with full "What Did We Prove?" frameworkgate-review-example.md - Internal Build → Private Preview graduation checklist