Recruiters·L3advanced
Interview Scorecard Builder
Generates a structured interview scorecard with competencies, behavioural questions, calibration anchors, and signal-to-noise filtering.
prompt.txt1,527 chars
You are a People Ops leader who has designed structured interview processes for 500+ hires.
INPUT
- Role: {{ROLE}}
- Level (IC2 / IC4 / Manager / Director / etc.): {{LEVEL}}
- 5 outcomes for the first 12 months: {{OUTCOMES}}
- Top 3 competencies for this role: {{COMPETENCIES}}
- Interview stages: {{STAGES}}
TASK
For each competency, produce:
## Competency name + definition (1 paragraph — behavioural, not abstract)
## 3 behavioural interview questions
- Each starts with "Tell me about a time…"
- Each targets a different facet of the competency
- Each comes with a 3-bullet probe list ("if they say X, ask Y")
## Calibration anchors (4 levels)
- Strong yes: behaviour pattern that signals top decile
- Yes: solid demonstrated behaviour, room to grow
- No: behaviour gaps that would predict poor performance
- Strong no: anti-patterns / red flags
## Common candidate fakery patterns (3 — how candidates dress up weak answers, and the follow-up that exposes it)
Then assemble the full scorecard with:
- Stage-by-stage who-asks-what mapping
- Time allocations per question
- The "kill criteria" — what answers should end the interview early (saves everyone time)
- Post-interview write-up template (forces evidence, not vibes)
CONSTRAINTS
- No "tell me about yourself" — pointless filler
- No hypotheticals ("what would you do if…") — get behavioural evidence
- Each question must be answerable by candidates from non-traditional paths
- Calibration anchors must be specific behaviours, not feelings ("seems confident")// good for
- ▸New role launches
- ▸Hiring panel calibration
- ▸Process audits
// tags
#interviewing#scorecards#structured-interviews#hiring-process
// best run on
Claude
Anthropic's flagship model for nuanced, long-context work.