Performance Review Writer

You're copying last year's review and hoping no one notices.

Performance Review Writer is an OpenClaw skill that ingests your 1:1 notes, peer feedback, and project summaries for a direct report, clusters the evidence into strengths, growth areas, and impact, drafts a 5-section review with cited examples, and runs an automatic bias and tone check before you submit. Unlike generic review prompts, it enforces a no-vague-praise rule: every strength needs a concrete example, or the script tells you exactly where the gap is.

What it does

• Synthesizes evidence from 1:1 notes, peer feedback, and project links into labeled clusters
• Drafts a 5-section review (impact, strengths, growth, calibration narrative, next-cycle goals) with source-attributed evidence
• Runs a bias and tone check flagging gendered language, vague praise, hedging, and calibration inconsistencies
• Maps the narrative to an engineering, PM, or design competency rubric for calibration alignment

What's in the package

All files listed below are bundled as fenced code blocks inside SKILL.md — there's no separate folder to install. Copy each code block into its own file using the path shown above it.

• SKILL.md — the full four-step workflow (gather, draft, check, calibrate) your agent follows.
• scripts/synthesize_evidence.py — clusters 1:1 notes, peer feedback, and project summaries into strengths/growth/impact with source attribution.
• scripts/draft_review.py — generates a 5-section review draft from the evidence clusters, with [ADD EXAMPLE] placeholders where evidence is thin.
• scripts/calibration_check.py — flags gendered language, vague praise, hedging, and calibration-inconsistent terms before submission.
• scripts/examples/ — sample notes, peer feedback, clusters, draft review, and calibration check output.
• data/competency-rubrics.yaml — IC1-Staff, associate-director PM, and junior-principal design rubrics with observable behaviors per level.

How it ships

Drop SKILL.md into ~/.openclaw/skills/performance-review-writer/. Restart OpenClaw. Run /performance-review-writer or say "Help me write Jordan's performance review" — your agent picks it up.

Pairs well with the Weekly 1:1 Prep skill — the notes you keep there become the evidence you synthesize here.

Questions: hackedcubeapps@gmail.com (response within 48 hours).