Test gap review

Skill yarlson/yarstack/plugins/yarstack/skills/test-gap-review

Engineering workflows and standards for Codex and Claude Code: plan, implement, test, review, and deliver repository changes.

Install

npx -y skills add yarlson/yarstack --skill test-gap-review

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Review whether existing tests and verification credibly prove a scoped behavior contract. Use during validation when evidence may omit important behavior or provide false confidence.

SKILL.md

1.9 KB, 304 tokens by cl100k_base, as published. Nobody here has run it

Test Gap Review

Audit verification sufficiency without silently becoming a test-writing or production implementation workflow.

Workflow

Establish the authoritative behavioral contract, comparison scope, and claimed verification.
Map invariants, assumptions, normal behavior, important boundaries, meaningful failures, preserved state, cleanup, and external effects to existing evidence.
Identify weak assertions, fixtures that cannot fail meaningfully, implementation-detail tests, unsupported manual claims, and skipped required commands. Flag example-only evidence when the claim spans a broad input space, operation sequence, concurrent ordering, restart, or recovery, and check that generated failures are reproducible.
Review changed test code as maintained product code. Flag unclear naming or structure, repeated setup that should use a focused helper, helpers that hide scenario data or mix responsibilities, and table-driven cases that do not share one execution path.
Report each material gap with the affected claim, missing evidence, consequence, smallest proving check, and residual risk.
If remediation is authorized, use test-design for test changes and behavior-implement for production behavior; otherwise remain read-only.

For each material claim report:

VERIFIED | NOT VERIFIED | INCONCLUSIVE
Claim: ...
Evidence: ...
Reasoning: ...

Do not chase coverage percentages, demand a stronger method without a concrete risk, add fake-confidence tests, broaden infrastructure, or change production behavior directly.

Finish when material claims have supported statuses and remaining verification risk is explicit.

What ships with it

Read from the repository