majiayu000/claude-skill-registry-data/assumption-tracker

Skill: assumption-tracker

What I do

I surface and manage hidden assumptions. I ensure that every leap of faith in a design or plan is documented, tiered by risk, and systematically validated through evidence or testing.

When to use me

• Before starting a new feature or architectural change
• When requirements are ambiguous or "common sense" is invoked
• During technical planning sessions to identify "we think" vs "we know"
• When evaluating third-party libraries or external API behaviours

Core principles

1. Surface the hidden — If it isn't proven, it's an assumption.
2. Tier by risk — Focus validation on assumptions with high impact and low certainty.
3. Validate early — Use spikes, prototypes, or data lookups to turn assumptions into facts.
4. Document outcomes — Record whether an assumption was proven true or false.

Patterns & examples

Assumption Logging Format:

Validation Techniques:

• Spike: Write a small, throwaway script to test a technical hypothesis.
• Prototype: Build a minimal UI to verify user interaction assumptions.
• Data Lookup: Query logs or databases to confirm usage patterns.

Anti-patterns to avoid

• ❌ "Trust me" logic — Relying on seniority instead of evidence.
• ❌ Validation lag — Building a full system on unverified, high-risk assumptions.
• ❌ Silent assumptions — Failing to voice doubts during the planning phase.

KB Reference

~/vaults/baphled/3. Resources/Knowledge Base/AI Development System/Skills/Thinking-Analysis/Assumption Tracker.md

Related skills

• critical-thinking — Rigorous analysis of claims
• epistemic-rigor — Distinguishing belief from knowledge
• prove-correctness — Evidence-based validation
• pre-action — Deliberate thinking before execution