From autocode
Converts problem_audit difficulty signals into a Codeforces-style rating (800-3500) with reasons, confidence, and calibration notes.
How this skill is triggered — by the user, by Claude, or both
Slash command
/autocode:problem-difficulty-ratingThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this after `problem_audit(include_difficulty=true)` has produced `difficulty_signals`.
Use this after problem_audit(include_difficulty=true) has produced difficulty_signals.
rating, band, and confidence from difficulty_signals.estimated_complexity, algorithm_tags, constraint_scale, implementation_evidence, and data_strength.risk_report.rating: CF-style integer from 800 to 3500.band: 入门 / 基础 / 中等 / 较难 / 困难 / 高难.confidence: low / medium / high, mapped from the numeric confidence.reasons: 3-5 bullets that cite concrete signals.why_not_lower: one sentence explaining why the problem is not easier.why_not_higher: one sentence explaining why the problem is not harder.calibration_notes: uncertainty and human-review notes.confidence < 0.5, mark the rating provisional and list missing evidence.problem_audit over intuition.npx claudepluginhub sztu-acm/autocodeDetails PluginEval's skill quality evaluation: 3 layers (static, LLM judge), 10 dimensions, rubrics, formulas, anti-patterns, badges. Use to interpret scores, improve triggering, calibrate thresholds.
Builds a scoring rubric interactively, evaluates an artifact with multiple models in parallel, then autonomously improves it one criterion at a time until a score threshold is met or circuit breaker fires.
Assesses code, designs, or approaches with 0-10 rating, pros/cons analysis, and actionable recommendations. Use for evaluating quality or trade-offs.