Aggregate per-skill eval verdicts into run-level scorecards with weighted category scores.
npx skills add https://github.com/jjackson/ace --skill opp-eval
Please help me install this Agent Skill. Skill: opp-eval Source: https://github.com/jjackson/ace/tree/main/skills/opp-eval Command: npx skills add https://github.com/jjackson/ace --skill opp-eval
Consolidates individual per-skill evaluation verdicts into a single run-level scorecard to reveal overall health of an opportunity and point to improvement areas.
Run opp-eval after an opportunity run to generate the run-level scorecard and recommendations.
High-intent search queries and answers about installing and using this skill.
You aggregate per-skill evaluation verdicts into a run-level scorecard by reading individual verdict files, computing weighted category scores, and handling missing data gracefully to output both machine-readable YAML and human-readable scorecards.
A run-level scorecard for opportunity evaluation consolidates individual per-skill verdicts to reveal overall opportunity health, identify the weakest dimensions, and surface improvement recommendations for governance reviews.
You generate machine-readable verdict YAML from opportunity evaluations by running an aggregation process across per-skill verdicts that computes weighted category scores and writes structured YAML files alongside human-readable scorecards.
Opportunity evaluation supports --quick, --deep, and --monitor modes, allowing you to aggregate per-skill verdicts across different analysis depths to produce both machine-readable and human-readable scorecards.
Missing data points are handled gracefully when computing weighted category scores, ensuring the aggregation process continues to produce a complete run-level scorecard without failing on incomplete per-skill verdict files.
jdrodriguezScore intake call transcripts against rubrics with evidence citations and coaching notes.
oozoofrogAudit Codex skill repositories for structure, frontmatter, and governance readiness.
OliverOuyangMonitor Guanyuan dashboard metrics and trigger alerts on rule violations.
dimagi-internalGrade partnership research artifacts against a four-dimension rubric and output verdict YAML.
