Performance Review Calibrator

Helps write performance reviews, spot bias, run calibration sessions and draft development plans.

Free plan: up to 10 private skills and 3 members, no credit card.

What this skill does

It provides review writing guidelines by rating, bias detection patterns, calibration facilitation guides, and development plan generation. Use for review writing, manager calibration, or performance conversations.

  • performance
  • reviews
  • feedback
  • calibration
  • ratings
  • development
  • HR
  • management

SKILL.md

Overview

Transform performance reviews from dreaded paperwork into meaningful development conversations. This skill encodes the frameworks for fair evaluation, effective writing, and calibration that drives growth—not just documentation.

When to Use This Skill

  • Review writing — Drafting performance evaluations
  • Calibration sessions — Facilitating rating alignment
  • Feedback delivery — Preparing for review conversations
  • Development planning — Creating growth plans

Rating Framework

Standard Rating Scale

Rating Label Definition Distribution
5 Exceptional Consistently exceeds all expectations ~5%
4 Exceeds Frequently exceeds expectations ~20%
3 Meets Consistently meets expectations ~50%
2 Developing Sometimes meets expectations ~20%
1 Below Does not meet expectations ~5%

Rating Calibration Questions

For each rating, answer:

  • What specific outcomes support this rating?
  • How does this compare to peers at same level?
  • Would another manager rate similarly with this evidence?

Review Writing Guidelines

Structure

SUMMARY (2-3 sentences)
Overall assessment and key themes

STRENGTHS (3-5 bullets)
Specific accomplishments with impact

DEVELOPMENT AREAS (2-3 bullets)
Growth opportunities with examples

KEY ACCOMPLISHMENTS (3-5 bullets)
Major achievements this period

GOALS FOR NEXT PERIOD (3-4 bullets)
Clear, measurable objectives

Writing by Rating

Rating Tone Focus
5 Celebratory Impact, stretch achievements
4 Positive Accomplishments, continued growth
3 Balanced Solid delivery, development areas
2 Constructive Specific gaps, support needed
1 Direct Clear expectations, consequences

Strong vs. Weak Examples

Strength - Weak:

"Great communicator"

Strength - Strong:

"Led weekly stakeholder updates that reduced escalations by 40%, consistently praised by leadership for clarity and proactive issue flagging"

Development - Weak:

"Needs to be more strategic"

Development - Strong:

"Opportunity to think beyond immediate deliverables to anticipate downstream impacts—for example, the Q3 launch timeline didn't account for legal review, causing delay"

Bias Detection Patterns

Common Biases

Bias Description How to Counter
Recency Overweighting recent events Review full period notes
Halo/Horn One trait colors all Evaluate dimensions separately
Similar-to-me Favoring similar people Focus on outcomes, not style
Central tendency Rating everyone average Force distribution
Leniency/Severity Skewing high or low Calibrate with peers

Bias Check Questions

Before finalizing, ask:

  • Am I comparing to job expectations or to other people?
  • What evidence supports each rating dimension?
  • Would I rate this person the same regardless of [demographic]?
  • Am I rating potential or actual performance?

Calibration Session Guide

Pre-Calibration Prep

Managers should bring:

  • Ratings with supporting evidence
  • Specific accomplishments for each person
  • Development areas with examples
  • Comparison to level expectations

Calibration Meeting Structure

  1. Overview (5 min) — Distribution expectations, process
  2. Exception Review (30 min) — Discuss 5s, 1s, and any 2s
  3. Boundary Cases (20 min) — Discuss 3/4 and 2/3 borderlines
  4. Distribution Check (5 min) — Review final distribution
  5. Action Items (5 min) — Communication, follow-ups

Discussion Questions

  • "What would this person need to do to move up a rating?"
  • "If we only had one 5 to give, would it be this person?"
  • "Is this a 2 because of the person or the situation?"

Development Plan Framework

SMART Goals

Element Question Example
Specific What exactly? "Improve presentation skills"
Measurable How will we know? "Deliver 3 exec presentations"
Achievable Is it realistic? Within capability with stretch
Relevant Why does it matter? Tied to career growth
Time-bound By when? "By end of Q2"

Development Actions

Development Need Actions
Technical skill gap Training, certification, projects
Leadership growth Stretch assignment, coaching, mentorship
Communication Feedback, practice opportunities
Strategic thinking Exposure to strategy, cross-functional work

Difficult Conversation Framework

Delivering Below-Expectations Feedback

  1. Set context: "This is a formal performance conversation"
  2. Be direct: State the rating and key reasons
  3. Provide evidence: Specific examples
  4. Listen: Allow response and questions
  5. Forward focus: Path to improvement
  6. Document: Written follow-up

Script Template

"I want to have a direct conversation about your performance this period. Your rating is [X], which means [definition]. The main reasons are [specific examples]. I know this is difficult to hear—what questions do you have?"

Resources

references/

  • rating-definitions.md — Detailed rating criteria
  • bias-guide.md — Complete bias detection

assets/

  • review-template.docx — Review document format
  • calibration-template.xlsx — Calibration tracking

A skill is a written procedure, not a prompt. Add it to your library and it is yours to edit in plain language.

In your Agentman library

A skill you keep, not a prompt you paste

  • Versioned, not pasted

    Edit a skill once and every agent using it follows the new version.

  • Shared with your team

    One library with per-skill access control and a record of who did what.

  • One connection

    A single MCP link works across Claude, ChatGPT, Cursor and Agentman agents.

How Agent Skills work
Questions

Using a skill

What is a skill?

A skill is a written procedure an AI assistant reads before it does a task: the steps, the judgement calls, the standards and what "done" looks like. It lives in a SKILL.md file, sometimes with reference files beside it. A prompt is typed again each time; a skill is versioned and runs the same way for everyone who uses it.

How do I use this skill in Claude, ChatGPT or Agentman?

Create a free Agentman account and add the skill to your library. One MCP connection then makes your library available in Claude, ChatGPT, Cursor and Agentman agents, so you set it up once. To see it work before you sign up, use "Try it in Claude" on this page, which opens a chat that reads the public SKILL.md.

Can I edit the skill after I add it?

Yes. Cloning a skill puts a copy in your workspace that is yours to edit in plain language. Skills are versioned: edit one once and every agent using it follows the new version. The free plan covers up to 10 private skills and 3 members, no credit card.

Make it yours

Add Performance Review Calibrator to your library, edit it to match how your team works, and use it from Claude, ChatGPT or an Agentman agent. The free plan covers up to 10 private skills and 3 members, no credit card.

Or try it in ChatGPT first.

Use this skill free