メニュー
Create new behavioral interview specs for the spec-driven-behavioral-interviews repository. Use when adding a new interview dataset, choosing category placement, scaffolding data/book or data/examples entries, filling the JSON schema, or preparing a new interview for the static explorer.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Add or improve Markdown bold emphasis for key phrases in behavioral interview JSON text. Use when highlighting competency briefs, prompts, prompt variants, interviewer intent, evidence, decision criteria, trade-offs, operating mechanisms, red flags, follow-up probes, or rubric signals with **key phrase** markers while preserving valid project data.
Promote a specific behavioral interview on LinkedIn from a provided interview.json URL or path. Use when drafting LinkedIn copy that explains behavioral interviews, their education and interviewing value, the project, and includes both the concrete interview anchor link and the full catalog link.
Review behavioral interview specs in this repository for realism, balance, educative value, interviewer usability, and data quality. Use when asked to critique interviews, audit the catalog, identify gaps, score interview quality, or write/update REVIEW.md with findings and recommendations.
Review behavioral interview specs in this repository for realism, balance, educative value, interviewer usability, and data quality. Use when asked to critique interviews, audit the catalog, identify gaps, score interview quality, or write/update REVIEW.md with findings and recommendations.
Improve existing behavioral interview specs in this repository. Use when upgrading an interview JSON file, making generic examples more concrete, strengthening strong and weak answer patterns, deepening follow-up probes, improving evaluation rubrics, or aligning an existing interview with the project schema and quality bar.
Add or improve concrete example stories, prompt variants, weak and strong answer variants, and follow-up angles for behavioral interview specs in this repository. Use when a user asks for more concrete examples inside each interview, more variants by role or level, or better example coverage.
| name | behavioral-interview-author |
| description | Create new behavioral interview specs for the spec-driven-behavioral-interviews repository. Use when adding a new interview dataset, choosing category placement, scaffolding data/book or data/examples entries, filling the JSON schema, or preparing a new interview for the static explorer. |
AGENTS.md, data/README.md, the target group manifest such as data/book/index.json, and one or two nearby interview.json files in the same competency family.book for canonical catalog content and examples only for small renderer examples.python3 _scripts/scaffold_interview.py --list-families
python3 _scripts/scaffold_interview.py --list-roles
python3 _scripts/scaffold_interview.py book new-interview-id \
--title "New Interview Title" \
--category-id judgment-strategy \
--category-name "Judgment and Strategy"
data/<group>/<id>/interview.json. Do not edit generated files under docs/.Situation -> Task -> Action -> Result -> Reflection -> System Change.Include these fields with concrete senior-leadership content:
roleFocus, difficulty, and competency with a clear family and definition.prompt.primary, at least two prompt.variants, and interviewer intent.scenario with a signature example, context, and constraints.storyAnatomy with personal accountability, trade-offs, measurable results, reflection, and durable operating change.exampleStories with distinct level, situation, strong answer sketch, why it works, weak version, and follow-up angles.strongAnswerPattern, weakAnswerPattern, followUps, evaluation.signals, evaluation.rubric, visualizations, practiceTemplate, and relatedScenarios.Make the interview test a real senior operating judgment, not a generic behavioral prompt. Strong content names the ambiguity, conflict, constraints, personal decisions, rejected alternatives, measurable outcomes, and mechanisms left behind. Avoid vague claims such as "aligned stakeholders" unless the spec explains the actual forum, decision rule, metric, or ownership model.
Run the focused checks before finishing:
python3 _scripts/validate_interviews.py
python3 _scripts/summarize_coverage.py
python3 build.py
python3 _scripts/smoke_static_site.py