HR Performance Review Expert

What Problem It Addresses

Performance reviews often suffer from inconsistent rating scales, vague goals, and scattered multi-source feedback. This skill is aimed at HR managers and team leads who need structured, explainable outputs from local Excel performance data, rather than a single aggregate score with no evidence.

How It Works

All analysis runs through scripts/run.py with a JSON action field:
- review: detects leniency, strictness, centrality, halo, and recency bias, then compares the employee against department and company benchmarks to produce a risk level, adjusted score, and calibration actions.
- goal: scores a goal on the Specific, Measurable, Achievable, Relevant, Time-bound dimensions, flags missing elements, and rewrites it into a SMART formulation.
- feedback: aggregates superior, peer, subordinate, and self 360-degree feedback into dimension sentiment, strengths, gaps, consensus, disagreement, and source-consistency risks.
- list: lists employees, company average score, and basic fields to confirm the data scope.

Boundaries and Caveats

The skill requires a local absolute Excel path, and the workbook should include name, employee ID, review period, evaluation dimensions, scores, and overall grade. It is not a replacement for HR policy or final performance decisions; it is a data-checking and bias-flagging layer. If an employee is missing, the path is absent, or the JSON is malformed, the caller must fix the input based on the script error. Files are read and processed locally and are not uploaded to external services.