Test Strategy
SkillDev toolsSelects the narrowest sufficient test boundary from requirements, repository evidence, and maintenance cost. Use when deciding integration or E2E coverage.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Test Strategy skill
What this skill tells your AI
The instructions your AI receives, as published by shinpr/agentic-code in .agents/skills/testing-strategy/SKILL.md and read by ahel’s review.
Purpose
Select tests that prove an observable requirement at the lowest sufficient cost. Candidate discovery does not create an obligation to implement every technically valid test.
Selection Process
For each acceptance criterion that may require integration or E2E proof:
- Name the observable boundary: Identify the interaction, persisted effect, process, or user journey that must be proven.
- Check existing proof: Reuse or extend an established test when it already observes the required behavior.
- Select the narrowest level:
- use a unit or local check for isolated behavior;
- use integration when the requirement depends on a named component, persistence, or process boundary;
- use E2E only when the complete journey is itself the requirement.
- Compare cost: Consider setup, execution time, brittleness, maintenance, and diagnostic value using repository evidence rather than fixed universal scores.
- Retain necessary proof: Select a test only when a cheaper boundary cannot establish the required behavior or protect an accepted contract.
No new integration or E2E test is a valid conclusion when existing or cheaper proof is sufficient.
Priority
Prioritize proof for:
- core user or consumer journeys;
- shared or public contracts;
- persisted state transitions;
- security, authorization, compliance, or irreversible effects;
- integrations with demonstrated failure risk.
Priority does not determine test level. Use the narrowest boundary that can observe the prioritized behavior.
Deduplication
Before adding a test:
- search the existing suite for equivalent behavior;
- identify overlapping proof at other levels;
- extend existing coverage when that remains clear and maintainable;
- keep one representative proof when several candidates observe the same risk.
Selection Record
Record only decisions needed by planning or review:
Acceptance criterion: [identifier or behavior]
Selected proof: existing | local | integration | E2E | none
Boundary observed: [observable interaction or result]
Reason: [why this is the narrowest sufficient proof]
Completion Check
- Every selected test traces to an observable requirement or accepted contract.
- A cheaper sufficient boundary was considered.
- Existing proof was reused where sufficient.
Signals
- GitHub stars
- 49
- Forks
- 5
- Last commit
- Aug 2026
Advanced
- Catalog kind
- skill
- Gateway key
testing-strategy-shinpr- Source
- github.com/shinpr/agentic-code