quality-vector-tests

Write vector-attributed test scenarios tracing PRD requirements to executable test suites.

10|9|Updated May 22, 2026
One-click install
npx skills add https://github.com/constructorfabric/insight --skill quality-vector-tests-constructorfabric
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: quality-vector-tests
Source: https://github.com/constructorfabric/insight/tree/main/.claude/skills/quality-vector-tests
Command: npx skills add https://github.com/constructorfabric/insight --skill quality-vector-tests-constructorfabric

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Feature specifications often lack traceable test coverage: requirements exist in PRDs, but there is no structured chain linking them through design allocations to concrete scenarios and executable tests. This Skill authors and reviews the Testing section of Insight FEATURE documents so every applicable FR/NFR has a scenario, a quality vector, a suite, and a linked test. ## Core Features & Use Cases - Requirement-to-test traceability: Builds the chain PRD FR/NFR → DESIGN allocation → FEATURE requirements → scenario → test, preserving canonical feature IDs and stable scenario numbers. - Vector and suite mapping: Assigns each scenario one of five quality vectors (Efficiency, Reliability, Performance, Security, Versatility) and the cheapest adequate test suite, from rust-unit and fe-unit up to stand-api, stand-ui, or observed telemetry. - Coverage gap review: Uses scripts/counts.sh to take repo-wide denominators (connectors, catalog metrics, dbt models, stand tests) that fail loudly as MOVED rather than reporting a misleading zero. - Use Case: When adding a Testing section to a new FEATURE for an ingestion connector, the Skill resolves inherited NFRs, drafts numbered scenarios with requirement citations, and links each implemented test back to its scenario. ## Quick Start Ask the assistant to write or review the Testing section of a FEATURE document, tracing its PRD requirements to vector-attributed scenarios and executable tests.

Frequently Asked Questions about quality-vector-tests

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

FAQPage Schema
How do I trace PRD requirements to tests in a feature spec?

Trace each FR/NFR from the PRD through the DESIGN allocation into the FEATURE's Requirements field, then write numbered scenarios citing those IDs and link each scenario's Test field to the exact implemented test file and function.

What are the five quality vectors for test scenarios?

The five vectors are Efficiency, Reliability, Performance, Security, and Versatility. Each scenario gets exactly one primary vector identifying its verification claim, with independent claims in other vectors handled by separate assertion-focused tests.

How do I choose the right test suite for a scenario?

Pick the cheapest existing suite capable of falsifying the claim: fe-unit before stand-ui for frontend, rust-unit before metric-spec or stand-api for serving, and connector-tests before dbt-tests for ingestion. Use observed telemetry for performance claims already measured in operation.

When should I use observed telemetry instead of running a test?

Use observed when operational telemetry already measures the claim on real traffic, such as latency percentiles or error rates. The entry must state the panel or query, time window, and environment, and it cannot gate a merge since it is lagging evidence.

What happens when a requirement has no testable scenario yet?

Record it as an unresolved decision in feature context with an owner and resolution point, not as a placeholder scenario. Missing infrastructure or unsettled product decisions are gaps to track, never reasons to mark a vector n/a or invent targets.