Agent skill
dev-inquiry
Developer inquiry skill for technical investigation, validation, and decision-making. Use when exploring unfamiliar technology, validating approaches with spikes, comparing options, or making architecture decisions. Covers the full thinking process from "I don't know" to confident decision.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/development/dev-inquiry
SKILL.md
Dev Inquiry
Feynman-style technical inquiry for developers. Understand before you decide. Validate before you build.
"The first principle is that you must not fool yourself—and you are the easiest person to fool." — Richard Feynman
Inquiry Modes
This skill supports four modes that often chain together:
| Mode | When to Use | Trigger Phrases |
|---|---|---|
| Investigate | Understand something unfamiliar | "explore", "understand", "how does X work" |
| Spike | Validate approach before building | "spike", "validate", "prove this works" |
| Compare | Evaluate options for your context | "compare", "vs", "which is better" |
| Decide | Make a concrete choice | "should we", "decide", "recommend" |
How to Use This Skill
- Identify the mode from the user's request
- Load the appropriate reference:
- Investigation → Read
references/investigation.md - Spike validation → Read
references/spike.md - Comparison → Read
references/scoring.md - Decision → Read
references/scoring.md(uses comparison + decision framework)
- Investigation → Read
- Follow the workflow in that reference
- Chain if needed — investigation often leads to spike, spike informs comparison
Mode Details
Investigate Mode
For understanding unfamiliar technology from first principles.
Process: Admit ignorance → Simplest experiment → Poke edges → Build mental model → Explain simply
Output: Mental model you can teach to someone else
Spike Mode
For validating technical feasibility before full implementation.
Process: Define scope (4-8 hours) → Write tests first → Implement minimal → Test with real data → Document pattern
Output: Proven pattern ready to replicate, or pivot decision
Compare Mode
For evaluating multiple options against your specific context.
Process: Define context → Choose criteria → Weight by importance → Score with evidence → Sanity check
Output: Weighted comparison matrix with evidence
Decide Mode
For making a concrete choice with documented reasoning.
Process: Ensure understanding (investigate if needed) → Compare options → State recommendation → Document tradeoffs → Assess reversibility
Output: Clear recommendation with rationale and risks
The Natural Flow
"Let's explore @Observable"
↓ Investigation
"Can I actually build nested observation?"
↓ Spike (validates understanding)
"@Observable vs @StateObject for my app"
↓ Comparison
"Should we adopt the new Observation framework?"
↓ Decision
Each mode builds on the previous. Don't decide before you understand. Don't compare before you investigate.
Quick Reference
| Request | Mode | Reference |
|---|---|---|
| "Let's explore Swift macros" | Investigate | references/investigation.md |
| "I want to understand async/await" | Investigate | references/investigation.md |
| "Spike SwiftData before we commit" | Spike | references/spike.md |
| "Validate this architecture works" | Spike | references/spike.md |
| "Compare REST vs GraphQL" | Compare | references/scoring.md |
| "SwiftData vs CoreData?" | Compare | references/scoring.md |
| "Should we use Combine?" | Decide | references/scoring.md |
| "Recommend an approach" | Decide | references/scoring.md |
Examples
For concrete examples across all modes, see references/examples.md.
Didn't find tool you were looking for?