Performance Review Prepper

Compiles your achievements, metrics, and impact stories from multiple sources into a polished, review-ready document. This skill pulls from git commit history, project notes, and raw bullet points to generate structured self-assessments with quantified impact. Works across review frameworks used by companies like Google, Amazon, Microsoft, and standard HR platforms like Workday and BambooHR.

Supported Platforms & Integrations

Platform	Integration Type	Features
GitHub / GitLab	Git history analysis	Commit counts, PR reviews, code impact metrics
Jira / Linear	Task board summary	Completed tickets, story points, sprint velocity
Workday	Review format export	Structured output matching Workday review forms
BambooHR	Review format export	Goal-aligned achievement mapping
Google Docs	Export format	Clean document with headers and bullets
Notion	Export format	Database-compatible achievement entries

When to Use This Skill

• Use this when your quarterly or annual review cycle is approaching and you need to compile achievements

• Use this when you cannot remember what you accomplished over the past 6 months and need a structured recall process

• Use this when your manager asks for a self-assessment and you struggle with self-promotion

• Use this when you want to build a "brag document" that you update incrementally throughout the year

• Use this when preparing for a promotion case and need to present impact across multiple projects

• Use this when switching managers and need to bring a new manager up to speed on your contributions

• Consider alternatives when you need to write peer feedback for a colleague (use a feedback writing skill)

• Consider alternatives when you need career coaching on what to pursue next (use a career planner)

Quick Start

Copy
# review-prep-config.yml
review:
  period: "2025-H2"
  framework: "standard"
  role: "Senior Software Engineer"
  level: "IC4"
  company_values:
    - "Customer Obsession"
    - "Ownership"
    - "Bias for Action"
  sources:
    - type: "git"
      repo: "./my-project"
    - type: "notes"
      file: "./achievements.md"

First run example:

> Prepare my H2 2025 performance review. I led the checkout redesign, mentored 2 junior devs, and reduced API latency by 40%.

Generated Review Document:
- Impact Summary (executive-level, 3 sentences)
- Key Achievements (5 bullets with metrics)
- Technical Contributions (PRs merged, code review stats)
- Leadership & Mentorship (specific examples)
- Alignment to Company Values (mapped to each value)
- Growth Areas & Goals (forward-looking)

Advanced Configuration

Copy
# Framework: Amazon (Leadership Principles)
amazon:
  leadership_principles: true
  pip_format: false
  narrative_style: "STAR"

# Framework: Google
google:
  peer_feedback_prep: true
  rating_scale: "googleyness"
  project_impact_tiers: true

Parameter	Type	Default	Description
period	string	"last-6-months"	Review period: Q1, H1, H2, annual, custom range
framework	string	"standard"	Options: standard, amazon-lp, google, microsoft, custom
role	string	""	Your current role title for context
level	string	""	Your level (IC3, L5, Senior, etc.) for calibration
tone	string	"confident"	Options: humble, confident, assertive
include_metrics	boolean	true	Pull quantified metrics from git and task sources
company_values	array	[]	Map achievements to specific company values
promotion_case	boolean	false	Frame achievements for promotion readiness
peer_feedback	boolean	false	Generate peer feedback request templates
brag_doc_mode	boolean	false	Incremental mode for ongoing achievement tracking
word_limit	number	1000	Target word count for the final document

Core Concepts

Concept	Description
STAR Format	Situation, Task, Action, Result — the gold standard for achievement bullets
Impact Quantification	Converting vague contributions into measurable business outcomes
Value Mapping	Aligning achievements to company-specific values or leadership principles
Brag Document	A running log of achievements updated weekly for easy review compilation
Calibration Framing	Presenting work at the level appropriate for your current or target level

+------------------+     +-------------------+     +------------------+
| Raw Sources      | --> | Achievement       | --> | Impact           |
| (git, notes,     |     | Extractor         |     | Quantifier       |
|  task boards)    |     | (identify wins)   |     | (add metrics)    |
+------------------+     +-------------------+     +------------------+
                                                          |
                                                          v
+------------------+     +-------------------+     +------------------+
| Review Document  | <-- | Value Mapper      | <-- | STAR Formatter   |
| (final output)   |     | (company align)   |     | (structure each) |
+------------------+     +-------------------+     +------------------+

Workflow Examples

Scenario 1: Standard Quarterly Review

Input: Raw notes: "shipped checkout v2, fixed 23 bugs, reviewed 45 PRs, onboarded new hire, improved test coverage from 67% to 89%"
Output:
- Led end-to-end delivery of Checkout V2, reducing cart abandonment by 12% and increasing conversion rates across 3 payment providers (Stripe, PayPal, Apple Pay)
- Improved codebase reliability by resolving 23 production bugs and increasing test coverage from 67% to 89%, resulting in 30% fewer incidents in Q4
- Strengthened team capacity by onboarding 1 new engineer (productive in 2 weeks vs 4-week avg) and reviewing 45 pull requests with thorough, educational feedback

Scenario 2: Amazon Leadership Principles Format

Input: "Built internal tool that saved ops team 10 hours/week. Pushed back on VP's feature request because data showed users didn't want it."
Output mapped to LPs:
- Customer Obsession: Used A/B test data to push back on a VP-sponsored feature, redirecting engineering effort to a customer-requested improvement that drove 15% engagement increase
- Invent and Simplify: Designed and built an internal automation tool reducing ops team workload by 10 hours/week, annualized savings of $25K

Scenario 3: Promotion Case

Input: "I want to make the case for promotion from IC4 to IC5. Show cross-team impact."
Output: Structured promotion packet with:
- Scope expansion evidence (cross-team projects)
- Technical leadership examples (design docs authored, architecture decisions)
- Mentorship impact (junior devs promoted/grown)
- Business impact at IC5 calibration level

Best Practices

Keep a weekly brag log. The single biggest mistake people make is trying to remember six months of work in one sitting. Spend five minutes every Friday jotting down what you accomplished. Use brag_doc_mode: true to maintain a running document that the skill can compile at review time.

Quantify relentlessly. "Improved performance" means nothing without numbers. Always ask yourself: how much faster, how many users affected, how much revenue impacted, how many hours saved. Even rough estimates like "reduced load time by approximately 40%" are far more powerful than qualitative descriptions.

Frame for your target level, not your current one. If you are pursuing a promotion, describe your work using the language and scope of the next level. IC5 work emphasizes cross-team impact and technical direction. IC4 work emphasizes execution and individual contribution.

Map to company values explicitly. Reviewers and calibration committees look for evidence aligned to company values. Do not make them infer the connection. State it directly: "Demonstrating Customer Obsession, I analyzed 200 support tickets to identify the top 3 pain points, then shipped fixes that reduced ticket volume by 35%."

Include growth areas proactively. Self-assessments that only list strengths appear lacking in self-awareness. Include 1-2 genuine growth areas with a specific plan. This shows maturity and often earns you more credit than another achievement bullet.

Common Issues

Problem: Git history shows too many trivial commits. The skill filters commits by impact — merge commits, feature branches, and PRs with descriptions are weighted higher than typo fixes. Set a minimum threshold with min_commit_significance: "medium" to skip noise. You can also provide a list of key project branches to focus the analysis.

Problem: Achievements sound like job descriptions. This happens when you describe responsibilities instead of results. The skill forces STAR formatting which naturally shifts language from "responsible for" to "achieved." If output still feels flat, add more context about the before-and-after state of what you worked on.

Problem: Unsure how to quantify non-engineering work. Mentorship, documentation, and process improvements are quantifiable too. Count the number of people mentored and their outcomes. Measure documentation by page views or reduced questions in Slack.

Privacy & Data Handling

• Local processing: All achievement data, git history analysis, and document generation happens locally within your Claude Code session. Nothing is sent to external review platforms.
• Data retention: Your review content exists only during the active session. The brag document mode saves to a local file you specify, giving you full control over storage.
• Export options: Output as markdown, plain text, or structured JSON. Can be copy-pasted directly into Workday, BambooHR, or Google Docs review forms.
• Sensitive data: Git history analysis reads only commit messages and PR titles from local repos — it does not access code content, private repos, or credentials.
• Manager visibility: Nothing generated by this skill is shared with your manager or HR system automatically. You control what to submit.