Technology Advisor

<!-- AUTO-GENERATED by export-skills.py — DO NOT EDIT -->
---
name: technology-advisor
description: Systematic technology selection using the STRIDE framework (Scope, Test reversibility, Requirements, Identify options, Deep evaluation, Execute). Use when selecting databases, frameworks, cloud providers, languages, or any technical stack decisions. Matches evaluation rigor to decision reversibility -- lightweight for 2-way doors, thorough for 1-way doors. Produces weighted decision matrices with sensitivity analysis.
tags: [technology-selection, decision-making, stride, evaluation, adr]
---

# Technology Advisor

Guide systematic technology selection using the STRIDE framework. Match evaluation rigor to decision reversibility.

## When to Use

- "Help me pick/choose/select [technology]"
- "Should I use X or Y?"
- "What [database/framework/cloud] should I use?"
- "Recommend a [technology type]"
- Validating or reviewing past technology choices

## Core Principle

**Match rigor to reversibility:**
- **2-way door** (reversible): 1-3 days, lightweight evaluation
- **1-way door** (irreversible): 1-2 weeks, full STRIDE process

## The STRIDE Framework

```
S - Scope the Decision
T - Test for Reversibility
R - Requirements Definition
I - Identify & Screen Options
D - Deep Evaluation
E - Execute & Validate
```

### S - Scope (15-30 min)

Questions to ask:
1. What technology are we selecting? (database, framework, cloud, etc.)
2. What problem does this solve?
3. Who are the stakeholders?
4. What's the time horizon? (6 months, 2 years, 5+ years)

### T - Test Reversibility (5 min)

Reversibility factors:

| Factor | Easy to reverse | Hard to reverse |
|---|---|---|
| Migration cost | <5% of effort | >25% of effort |
| Data portability | Easy export | Proprietary lock-in |
| Team retraining | <1 week | >1 month |
| Time to switch | <2 weeks | >2 months |

Assessment:
- **2-way door** -> "Reversible. Use lightweight 3-day evaluation (Steps S,T,R,E)"
- **1-way door** -> "Hard to reverse. Use full STRIDE (2 weeks)"

### R - Requirements (1-2 hours)

**COMPRIS -- 7 universal criteria:**

| # | Criterion | Description |
|---|---|---|
| C | Cost | Total cost of ownership |
| O | Operational | Deployment complexity, support burden |
| M | Maintainability | Documentation, debugging, code quality |
| P | Performance | Speed, scalability, resource efficiency |
| R | Risk | Vendor stability, security, compliance, lock-in |
| I | Integration | API compatibility with existing systems |
| S | Skills | Team expertise, learning curve, hiring availability |

Weight criteria by normalizing 1-5 importance ratings to percentages (sum = 100%).

### I - Identify & Screen (2-4 hours)

1. **Generate options** -- research 10-20 initial candidates
2. **Apply hard filters** -- budget ceiling, platform support, compliance, maturity
3. **Present shortlist** -- narrow to 3-5 finalists

### D - Deep Evaluation (4-8 hours)

Score each option 1-5 per criterion:
- **5** = Excellent / Far exceeds
- **4** = Good / Exceeds
- **3** = Average / Meets
- **2** = Below average / Partially meets
- **1** = Poor / Does not meet

**Decision matrix:**

| Criterion | Weight | Option A | Option B | Option C |
|---|---|---|---|---|
| Cost | 25% | 4 (1.00) | 3 (0.75) | 5 (1.25) |
| Performance | 20% | 5 (1.00) | 4 (0.80) | 3 (0.60) |
| Skills | 15% | 3 (0.45) | 5 (0.75) | 2 (0.30) |
| **TOTAL** | **100%** | **3.45** | **3.30** | **3.15** |

**Sensitivity analysis:** Test if winner changes with different weight distributions or +/-1 score variations on uncertain criteria.

### E - Execute & Validate

Document the decision:
- Problem statement
- Shortlist evaluated
- Criteria & weights
- Scores & rationale
- Final recommendation
- Risks & mitigations
- Success metrics
- Review date

## Interaction Patterns

**Full guidance:** Walk through all 6 STRIDE steps for complex decisions.

**Quick assist:** User provides context upfront. Focus on criteria weights, score options, present recommendation.

**Second opinion:** User has a choice, needs validation. Review their logic, test sensitivity, identify blind spots.

**Decision review:** Reassess a past decision against original success metrics.

## Anti-Patterns to Prevent

| Pattern | Response |
|---|---|
| **Analysis paralysis** | "We've been evaluating 2 weeks on a reversible decision. Pick one, validate with 3-day prototype." |
| **Premature commitment** | "This is a 1-way door. Spend 2 weeks evaluating, not 2 days." |
| **Feature-list syndrome** | "Both have 90% of features. Focus on the 3 criteria that differentiate them." |
| **Recency bias** | "HackerNews loves [new tech] but it has 500 stars. [Mature option] has 50K stars and proven stability." |
| **Sunk cost fallacy** | "2 months invested, but signals say it's wrong. Migration is 3 weeks. Sunk cost is not a reason to continue." |

## Key Principles

1. **Match rigor to reversibility** -- don't over-analyze reversible choices
2. **Requirements before options** -- define needs before researching solutions
3. **Weights reveal priorities** -- if user can't weight criteria, priorities are unclear
4. **Data over opinions** -- score based on benchmarks, not marketing
5. **Document decisions** -- future you needs to know why you chose this
6. **Set review dates** -- technology landscape changes, reassess periodically

## References

- Source: claude-skills/technology-advisor (MIT License)
- STRIDE framework for systematic technology evaluation

<!-- Source: .faos/custom/skills/architecture/technology-advisor/SKILL.md -->