Analyze job grading outputs across seven axes for structural consistency.
npx skills add https://github.com/Capataina/Cernio --skill test-grade-jobs
Please help me install this Agent Skill. Skill: test-grade-jobs Source: https://github.com/Capataina/Cernio/tree/main/.claude/skills/test-grade-jobs Command: npx skills add https://github.com/Capataina/Cernio --skill test-grade-jobs
This Skill addresses the challenge of maintaining grading consistency in a job-discovery engine by measuring how agents apply a rubric across a stratified sample of job listings, ensuring the grading logic remains stable and structurally sound over time.
Invoke the test-grade-jobs skill to perform a full rubric consistency audit and generate a report on the current grading quality.
High-intent search queries and answers about installing and using this skill.
To audit grading consistency, this Skill measures the structural integrity of job grading workflows through a multi-axis analysis of agent outputs, evaluating format adherence, reasoning specificity, and internal consistency across stratified samples.
Rubric validation detects regressions by persisting grading scores to a baseline, allowing you to automatically track if rubric changes or agent updates introduce structural drift, unintended bias, or degraded output quality over time.
Structural auditing evaluates grading outputs across seven distinct axes, including format adherence and reasoning specificity, ensuring that automated agents apply rubrics coherently and maintain inter-agent consistency across job listings.
No external dependencies are required. The Skill performs stratified sampling and statistical analysis of grading metrics using only the Python 3 standard library, ensuring a zero-dependency setup for your CI/CD pipelines.
You should run a rubric consistency audit whenever you update the grading rubric or modify agents, verifying that the changes produce coherent grades without introducing unintended bias or structural drift in the job-discovery engine.
Yes. The Skill applies directly to CI/CD pipelines for quality assurance of automated career-preparation engines, automatically persisting scores to a baseline to detect if rubric changes cause performance regressions in grading workflows.
andikarachmanAudit seeds, versions, data hashes, and environments for ML experiment reproducibility.
Ankh-StudioRun automated regression tests comparing evaluations against baseline metrics.
softmgAutomate GitHub Actions workflow generation and CI/CD pipeline management.
PeterSalvatoDiagnose coherence across Governance, Infrastructure, and Output tiers of a website.
