Interviewer Calibration · 30 Minutes Before the Loop

You're my interview-loop calibration partner. I'm the recruiter running a 5-round loop. Build a 30-minute calibration brief for the panel.

# Context
- Role + level: {position_level}
- Role thesis (what this person actually solves): {role_thesis}
- 5 interviewers + roles + which signal each tests: {panel_assignments}
- Candidate pipeline (finalists, each one's hot spot): {candidate_pipeline}

# Output three sections

## 1. 5 must-test signals + rubric (60 words each)
Based on role thesis. For each:
- Signal name (specific behavior, not 'ownership')
- Strong signal behavior (one concrete line)
- Mid signal behavior (one concrete line)
- No signal behavior (one line + one dark signal we'd be alarmed to see)

## 2. 5 calibration questions for the panel (50 words each)
One per interviewer. Pose a hypothetical candidate answer; ask the panel to pre-score:
- "If a candidate said this, would you give strong / mid / no?"
- Use the spread as calibration. If 3 say strong and 2 say mid, rubric isn't clear enough.

## 3. 5 debrief anti-patterns + counters (30 words each)
- e.g. "founder speaks first, anchors everyone" → "10 minutes of independent scoring first"
- e.g. "comparing candidates" → "evaluate against the bar, not each other"
- At least one about bias (school / gender / articulation vs substance)

# Boundaries
- Every signal must be testable (paired with a specific behavior). No "good leadership".
- Calibration answers must be concrete. No "it depends".