skills.rest
yakeworld
yakeworldCOMMUNITY
Explore all skills by this author →
GitHub

golden-test-methodology

Measures golden test coverage across skills and defines test creation methodology.

6|1|Updated May 11, 2026
Search Tags:#quality assurance#verification#test coverage#test methodology#skill audit#coverage report#golden test
One-click install
npx skills add https://github.com/yakeworld/Synthos --skill golden-test-methodology-yakeworld
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: golden-test-methodology
Source: https://github.com/yakeworld/Synthos/tree/main/skills/private/extended/research-tools/synthos-akne-bridge/quality/golden-test-methodology
Command: npx skills add https://github.com/yakeworld/Synthos --skill golden-test-methodology-yakeworld

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams maintaining many AI skills lack a consistent way to measure whether each skill has complete golden tests, making quality gaps invisible until failures occur. ## Core Features & Use Cases - Coverage Measurement: Computes golden test coverage as the ratio of skills with complete golden sets (GOLDEN_SET.md plus cases/ plus expected/) to total skills, with graded thresholds from Critical to Excellent. - Methodology Definition: Defines structural rules for creating golden tests, including one-to-one case/expected pairing, weight differentiation, and synchronization after signature changes. - Prioritized Expansion: Orders gap-filling work by newly absorbed skills first, then high-usage skills, core atoms, and extension skills ranked by DIAGNOSE score. - Use Case: During an evolution cycle's DIAGNOSE step, run a coverage audit across all skills to flag those below 30% coverage as P0 systemic quality gaps. ## Quick Start Audit the golden test coverage of all skills in the repository and report the coverage grade, gaps, and prioritized next targets.

Frequently Asked Questions about golden-test-methodology

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

FAQPage Schema
How do I measure golden test coverage across multiple skills?▼

Count a skill as covered only when golden/GOLDEN_SET.md, golden/cases/, and golden/expected/ all exist and are non-empty, then divide covered skills by total skills. Coverage at or above 70% is acceptable, while below 30% indicates a systemic quality gap.

What makes a valid golden test set for a skill?▼

A valid golden set requires three components: a GOLDEN_SET.md definition document, a cases/ directory with test inputs, and an expected/ directory with matching expected outputs. Cases and expected files must pair one-to-one by name, and weights must not all be critical.

Does an empty golden directory count toward test coverage?▼

No. A golden/ directory that exists without actual cases and expected files is treated as having no tests and is excluded from coverage statistics. All three components must exist and be non-empty.

When should golden tests be updated after skill changes?▼

Golden tests must be updated whenever a skill's signature or output format changes. Stale tests validating outdated behavior produce unreliable verification results and undermine the quality gate.

Which skills should get golden tests first when expanding coverage?▼

Newly absorbed skills get golden tests immediately upon absorption, followed by high-usage skills ordered by frequency. Core atoms are maintained continuously, and extension skills are addressed from lowest to highest DIAGNOSE score.

Related Skills

Explore More Software Engineering Skills→
agentxmagentxm
5

another-skill

Enumerate skills in a repository by parsing SKILL.md frontmatter.

Official
Intermediate
BIGSHOLBIGSHOL

manage-skills

Analyze session file changes to identify verification skill coverage gaps.

Community
Advanced
makeupmakeup
47

add-unit-tests

Write missing unit tests for makeup-js modules using BDD and vi conventions.

Official
Advanced
Sir-chawakornSir-chawakorn
9

property-based-testing

Generate diverse inputs and shrink counterexamples to verify software invariants.

Community
Advanced