Panel Synthesizer

Hiring Panel Feedback Synthesizer

When to use this skill

A candidate has been interviewed by a panel of 3 or more people. Each panelist has written their own feedback. The hiring manager needs to synthesize this into a single hire/no-hire recommendation. Done casually, this process favors the loudest panelist and under-weights the quiet skeptic. This skill is calibrated to do it correctly.

Input the user provides

• Candidate name
• Role they are being considered for
• Panelist write-ups (3 to 7 typical)
• Optionally: the rubric or competencies the panel was supposed to evaluate

What this skill does

1. Reads each panelist's feedback in full
2. Identifies the collective signal across all panelists
3. Specifically surfaces:

• Areas of strong agreement (both positive and negative)
• Areas of disagreement (and which panelists are on each side)
• The most specific concern raised (often the most important signal, regardless of how loudly it was raised)
• The hidden veto (a panelist saying "strong overall, but..." is often a soft no, calibrate accordingly)

1. Produces a one-page decision memo in the structure below

Output structure

# Candidate: [Name]
*Role: [role]. Panel size: [n]. Interview dates: [range].*

## Recommended decision
[Hire / No-hire / Conditional hire pending further conversation]

## Confidence in this recommendation
[High / Medium / Low], with one-sentence rationale.

## Where the panel agrees
- [Strength 1, cited by N panelists with specific evidence]
- [Strength 2]
- [Concern 1, if there is consensus on a concern]
- [Concern 2]

## Where the panel disagrees
- [Disagreement 1]: [Panelist A] thinks [X]. [Panelist B] thinks [Y]. The deciding question is [Z].
- [Disagreement 2]: same

## The most specific concern raised
[Quote or close paraphrase of the most specific, evidence-backed concern from any panelist. This is often the signal that matters most regardless of how loudly it was raised.]

## The hidden veto check
- Is any panelist saying "strong but" / "I would not block" / "leaning yes but"? If yes, name them and quote the qualifying phrase.
- What is the actual concern under the polite phrasing?

## What would change this recommendation
2-3 specific things, including any further conversations or work product that would resolve the open questions.

Calibration notes

• Weight specificity over volume. A panelist who provides a concrete example of a concern carries more signal than a panelist who provides three superlatives.
• The "hidden veto" check is the highest-leverage section. Most panels lose good no-hire signals because one polite panelist's qualifying language gets read as a yes. This skill forces explicit acknowledgment of the soft no.
• Quote panelists where useful. Their actual words are better than paraphrase.
• If panelists' write-ups are short or contain no evidence, flag this. "Strong communicator" is not feedback. The skill should note when feedback quality itself is the limiting factor.
• Do not invent agreement that does not exist. If the panel is genuinely split, say so and propose a path to resolve.

Bias awareness

When synthesizing, flag any of the following if they appear:

• Concerns that are phrased in coded language (e.g., "culture fit" without specifics)
• Strengths attributed only to surface traits (confidence, communication style) without skill evidence
• Disagreement that maps to demographic axes among the panelists

These are not automatic disqualifiers, but they warrant explicit examination in the memo.

When NOT to use this skill

• Single-interviewer feedback (no panel to synthesize)
• Reference checks (different signal structure)
• Performance reviews of existing employees (different rubric)

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: Kotrotsos
• Source: Kotrotsos/claude-skills-non-coding
• License: MIT

Install and usage instructions live in the source repository linked above.