property-based-testing_skill

Overview

This skill helps you apply property-based testing (PBT) when writing tests for serialization, validation, normalization, and pure functions. It provides a compact property catalog, pattern detection guidance, library references, and practical input and settings best practices. Use it to find the strongest meaningful properties and avoid common PBT pitfalls.

How this skill works

The skill inspects function signatures, naming patterns (encode/decode, normalize, sort, validate), and test intent to recommend appropriate properties (roundtrip, idempotence, invariant, commutativity, etc.). It suggests input strategies, Hypothesis/fast-check snippets, and test settings for development, CI, and nightly runs. It flags red flags like tautological assertions, vacuous tests, overuse of assume(), and accidental reimplementation.

When to use it

• Testing serialization/deserialization or encode/decode pairs (roundtrip)
• Verifying normalizers, sanitizers, or canonicalizers (idempotence)
• Checking pure functions with algebraic contracts (commutativity, associativity)
• Validating validators/normalizers where output must satisfy an invariant
• Testing data structure operations (add/remove/get invariants)

Best practices

• Constrain inputs in the strategy, avoid heavy use of assume()
• Limit sizes (max_size, max_length) to prevent slow tests
• Reuse shared strategies for consistent, maintainable tests
• Prefer the strongest applicable property (roundtrip > idempotence)
• Add explicit edge-case @example tests alongside generated cases

Example use cases

• Roundtrip test for JSON: decode(encode(x)) == x for domain types
• Idempotence for a normalize() function: normalize(normalize(x)) == normalize(x)
• Invariant for collections: remove(add(coll, x), x) leaves coll unchanged
• Oracle test: new_impl(x) == reference_impl(x) across generated inputs
• No-crash baseline for inputs that must never raise exceptions