shift-left-testing

Refines backlog Story acceptance criteria and drafts pre-sprint ATP outlines in Jira.

Updated Aug 31, 2026
One-click install
npx skills add https://github.com/ahuaracab/test-genesis --skill shift-left-testing-ahuaracab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: shift-left-testing
Source: https://github.com/ahuaracab/test-genesis/tree/main/.agents/skills/shift-left-testing
Command: npx skills add https://github.com/ahuaracab/test-genesis --skill shift-left-testing-ahuaracab

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Ambiguous or incomplete acceptance criteria cause defects to be discovered only after implementation. This Skill runs a pre-sprint QA refinement pass over a batch of backlog Stories, surfacing ambiguities, gaps, and edge cases before sprint planning so PO and Dev can fix requirements instead of production code. ## Core Features & Use Cases - Batch backlog triage: Resolves candidate Stories from explicit IDs or a JQL query, rejects non-Story types, applies veto rules and a risk score, and presents a ranked candidate table for user approval. - Structured AC refinement: Reuses sprint-testing's acceptance-test-planning phases to produce refined Given/When/Then scenarios, outline-only test design with coverage estimates, and PO/Dev question lists, all marked NEEDS PO/DEV CONFIRMATION where inferred. - Jira handoff automation: Writes refined ACs to the acceptance_criteria field, populates the acceptance_test_plan field, posts a handoff comment, adds shift-left-reviewed labels, transitions Stories to Estimation, and closes a [QA] Shift-Left Review tracking subtask. - Use Case: Paste a list like UPEX-100,101,102 sitting in Backlog and ask to groom them; the Skill triages, refines each Story sequentially, publishes results to Jira, and produces a batch report for sprint planning. ## Quick Start Ask the assistant to run shift-left refinement on a comma-separated list of backlog Story IDs such as UPEX-100, UPEX-101, and UPEX-102.

Frequently Asked Questions about shift-left-testing

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

FAQPage Schema
How do I refine acceptance criteria for backlog stories before sprint planning?

Pass a comma-separated list of Story IDs or ask to groom the backlog. The Skill triages candidates with veto rules and a risk score, refines each accepted Story into Given/When/Then scenarios plus outline-only test design, then publishes everything to Jira fields and transitions Stories to Estimation.

What is the difference between shift-left-testing and sprint-testing?

Shift-left-testing runs pre-sprint on batches of backlog Stories and produces refined ACs plus outline-only ATPs with no execution. Sprint-testing runs in-sprint per ticket from Ready For QA status, performing full test design, execution, and evidence capture.

Can shift-left refinement handle bugs or tech-debt tickets?

No. The Skill accepts only Stories because bugs are reactive and have no upstream acceptance criteria to refine. Non-Story types like Bugs, Spikes, and Tech-debt are rejected loudly with a one-line reason surfaced to the user.

Does this skill create Test Plan items in Xray or Jira?

No. The pre-sprint ATP is written to the acceptance_test_plan custom field, or a fallback comment when the field is absent. The actual Test Plan item is created later by sprint-testing Stage 1 from that field content once the Story enters a sprint.

What happens if a story was already shift-left reviewed?

Stories carrying the shift-left-reviewed label with a dated label less than 30 days old are surfaced separately as already refined. The user decides per Story whether to skip or re-refine, and stale refinements older than 30 days default to refresh.

Why does the skill stop transitions at the Estimation status?

The transition guardrail stops at Estimation because moving a Story to Ready For Dev is a PO and Dev lead decision after estimation. Stories already past Estimation keep the refinement content but skip workflow transitions with a logged warning.