generate-tests

Generates adversarial test suites for verified competitive programming problem solutions.

1|Updated Jul 17, 2026
One-click install
npx skills add https://github.com/vaskoyudha/agent-skills-problem-gen --skill generate-tests-vaskoyudha
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: generate-tests
Source: https://github.com/vaskoyudha/agent-skills-problem-gen/tree/main/.qwen/skills/generate-tests
Command: npx skills add https://github.com/vaskoyudha/agent-skills-problem-gen --skill generate-tests-vaskoyudha

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Competitive programming problems often ship with weak test suites that let incorrect solutions pass. This Skill produces a rigorous, adversarial test suite that deliberately targets every known wrong approach, ensuring only correct solutions survive. ## Core Features & Use Cases - Adversarial Test Design: Creates one targeted test per wrong approach listed in the solution, with documented failure mechanisms. - Systematic Edge Case Coverage: Applies a taxonomy covering arrays, graphs, strings, DP, trees, and math problems to guarantee boundary conditions are tested. - Stress Test Configuration: Defines brute-force-vs-reference comparison setups with 100+ random tests and valid input ranges. - Use Case: After a solution for a binary search problem is verified solvable, use this Skill to produce a test_suite.json with 15-30 cases across basic, edge, adversarial, boundary, and stress categories, plus a coverage report mapping each test to the wrong approach it breaks. ## Quick Start Generate a complete adversarial test suite for the verified solution in solution.json, covering every wrong approach and edge case category.

Frequently Asked Questions about generate-tests

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate adversarial test cases for competitive programming problems?▼

Read each wrong approach from the solution's common_wrong_approaches list, design a constraint-valid input that makes it fail, and compute the correct expected output via brute force. Document which approach each test breaks in the purpose field.

What test categories should a programming contest test suite include?▼

A complete suite needs five categories: basic tests including all samples, edge cases from a systematic taxonomy, adversarial tests targeting wrong approaches, boundary tests at constraint limits, and a stress test configuration with at least 100 random tests.

How many test cases does a competitive programming problem need?▼

Aim for 15-30 hand-written cases with a minimum of 13: 3-5 basic, 5-8 edge cases, 3-5 adversarial, and 2-3 boundary tests. Every wrong approach listed in the solution must have at least one dedicated adversarial test.

What happens if the solution fails solvability verification?▼

If the solution's solvability_verdict is SOLVABILITY_FAILURE, no tests are generated. The Skill outputs an empty test suite with a coverage report explaining the failure instead of fabricating tests.

Why do auto-generated test suites let wrong solutions pass?▼

Weak suites only test happy-path inputs and never target specific wrong approaches like greedy failures, off-by-one errors, or integer overflow. Effective suites design one input per known wrong approach and verify expected outputs with brute force.