Use when opening a QA, SDET, automation, test lead or quality manager role or running any part of the hiring loop: job description, interview questions, a competency-anchored scoring rubric, interviewer calibration, the panel debrief to a hire or no-hire decision and a 30-60-90 day onboarding plan.
中文简介正在准备中。当前显示原文说明。
实测评分 未实测
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Hiring for QA roles is calibration-heavy: the same question scored by two interviewers without a rubric produces high noise. The remedy, per the structured-interview research (Levashina et al. 2014, Personnel Psychology), is a structured interview - same questions, same order, same scoring rubric, calibrated interviewers. This kit runs the whole chain as one workflow, with one competency vocabulary (drawn from the ISTQB Foundation Level v4.0 syllabus) threaded through every artifact:
| Stage | Artifact | Deep reference |
|---|---|---|
| 1. Open the role | Job description + recruiter screening note | references/jd-author.md |
| 2. Ask | Role + seniority question bank (STAR, Bloom's K1-K4) | references/interview-questions.md |
| 3. Score | 4-level competency-anchored rubric | references/rubric.md |
| 4. Calibrate | Gold-standard answers + panel session script | references/calibration.md |
| 5. Decide | Panel debrief to hire / no-hire (workflow below) | this file, "Running the debrief" |
| 6. Ramp | 30-60-90 day onboarding plan | references/onboarding.md |
Stages 2-4 are the structured-interview tripod - what we ask, how to score, what each score looks like. None of the three is sufficient alone; running structured interviews requires all three.
Do not use this kit to:
skill-matrix-author (this plugin); its gap report is the ideal "why this
role is open" input to stage 1.Lock artifacts at the start of a hiring round; mid-round changes invalidate prior candidates' scores.
After the final interview, the panel converges on a decision in a facilitated calibration loop. Required inputs: role + seniority, the filled-in rubric with each interviewer's per-dimension scores, the calibration guide, and the panel list. Refuse to start if any interviewer's scores were not submitted independently before the debrief - per anchoring-bias research (https://en.wikipedia.org/wiki/Anchoring_effect), the first opinion revealed in a group setting has disproportionate influence on subsequent scorers; independent pre-submission is the primary structural mitigation.
INCOMPLETE_PANEL_SUBMISSION with the missing interviewer's
name and a deadline.MISSING_RUBRIC - halt);
that validity loss is what structured interviews exist to prevent
(Levashina et al. 2014,
https://en.wikipedia.org/wiki/Structured_interview).| Anti-pattern | Why it fails | Fix |
|---|---|---|
| Running interviews with questions but no rubric (or vice versa) | Half a structured interview still drifts; the pair is the unit. | Stages 2 + 3 travel together; stage 4 before the first candidate. |
| JD vocabulary differing from the rubric | Candidate applies to one role, gets scored on another; debriefs derail. | Stage 1 reuses the rubric's competency dimensions. |
| Skipping the calibration session | Inter-rater agreement does not appear without practice. | The stage-4 session is mandatory before the first real candidate. |
| Deciding by committee re-vote on a borderline | Reverts to unstructured group dynamics. | Escalate BORDERLINE to the hiring manager with the decision document. |
| Shelving the rubric after the offer | The hire's weak axes are known and then ignored. | Stage 6 converts borderline axes into targeted phase-2 development. |
skill-matrix-author (this
plugin) - its capability-gap report feeds stage 1.还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Use when a project has accessibility debt and a zero-violations gate would block every PR. Builds a CI ratchet that fails only on new WCAG violations a PR introduces, with pre-existing ones baselined per rule and page, from axe-core, pa11y, Lighthouse, WAVE or IBM Equal Access results.
中文简介正在准备中。当前显示原文说明。
Use when launching, auditing or governing an A/B test or online experiment design. Builds a per-experiment checklist and sign-off: pre-registered metric, power and guardrails, randomization unit and sample ratio mismatch checks, assignment, telemetry, peeking rules and novelty effects.
中文简介正在准备中。当前显示原文说明。
Use when a PWA's manifest, icons or install handler change, install conversions drop, or an install test is flaky. Builds Playwright tests for the install flow: installability criteria, the beforeinstallprompt and userChoice chain, the iOS Safari path and display-mode standalone after install.
中文简介正在准备中。当前显示原文说明。
Use when a spec and a test suite both exist but nobody can say which requirements are covered, before a release sign-off, an audit or planning test effort. Uses an LLM to map existing tests to spec sections, outputs a coverage matrix per spec ID and names the gaps. It does not write new tests.
中文简介正在准备中。当前显示原文说明。
Use when tests should be generated from acceptance criteria, user stories or requirements with an LLM. Scores confidence per test case (self-assessment plus assertion, naming and completeness checks), batches uncertain cases for human review and emits tests in the team's existing framework.
中文简介正在准备中。当前显示原文说明。
Use when a team wants richer test reports than JUnit XML: steps, attachments, per-test history, retries and severity or feature labels. Sets up Allure Report adapters for pytest, Jest, JUnit, TestNG, NUnit or Mocha, failure categories, history trends and the static HTML site as a CI artifact.
中文简介正在准备中。当前显示原文说明。