Extracts testable behaviors from specifications into a numbered T-ID checklist ledger.
npx skills add https://github.com/Hakkadaikon/hymme --skill test-extract-hakkadaikon
Please help me install this Agent Skill. Skill: test-extract Source: https://github.com/Hakkadaikon/hymme/tree/main/skills/test-extract Command: npx skills add https://github.com/Hakkadaikon/hymme --skill test-extract-hakkadaikon
tasks/test-design/<target>.md. - EARS-based classification: Labels each behavior as event, state, ubiquitous, unwanted, optional, boundary, or nonfunctional, forcing at least one abnormal-case row per normal-case row. - Bidirectional test review: Maps existing tests against the ledger in both directions to find missing tests, unjustified tests, and skipped tests that are silently not running. - Use Case: Before implementing tests for a payment module, run the extraction to produce a T-ID ledger covering double-payment prevention, expired-deadline operations, and boundary conditions, then hand the closed ledger to the test-catalog step for technique assignment. ## Quick Start Ask the AI to extract all testable behaviors from the target feature into a T-ID ledger using the test-extract skill.High-intent search queries and answers about installing and using this skill.
Define ambiguous terms first, then scan the target using anchors like acceptance criteria, function signatures, branches, or state transitions. Record each behavior as a numbered T-ID row in a checklist ledger under tasks/test-design/, classifying it with EARS labels such as event, state, or unwanted.
Build the behavior ledger first, then map existing tests onto its right side. Empty left-side rows reveal missing tests, tests without a matching behavior row are unjustified or brittle, and skipped tests count as unprotected even if listed.
EARS classifies requirements into event, state, ubiquitous, unwanted, and optional patterns. In test extraction it acts as a lightweight checklist forcing each normal-case behavior to have at least one corresponding abnormal-case row, catching boundary and error paths.
Use the loop-engineering route with TLA+ when the target involves state transitions, concurrency, or protocols requiring exhaustive state-space verification. Manual extraction suits typical features where systematic enumeration of behaviors is sufficient.
The gate requires the ledger to be fully checked with no missing T-ID numbers before any test is written, verified by grep commands on the ledger file. This prevents implementation from starting on an incomplete behavior inventory, which is the main source of coverage gaps.
Adgmed2018Generate Gherkin feature files from BR-FUTURE-NNN mappings for acceptance criteria.
latestaiagentsDesign test cases and suites using markdown and BDD/Gherkin formats.
Q16GOrchestrate gray-box testing by combining white-box candidate discovery with black-box request validation.
nrdxpProvides mathematical foundations and formal verification vocabulary for protocol and state machine design.