Execute agent-centric reviews on all rules in rules/ directory and generate prioritized improvement report
Produces best-in-class implementation plans with 15 mandatory sections covering architecture, final-state artifacts, dependency deltas, parity tables, test strategy, CI matrix, risk register, phased tasks with time estimates, rollback, acceptance criteria, and open questions. Use when authoring a migration plan, feature design, refactor plan, or any non-trivial implementation document. Triggers on "create a plan", "write a plan", "migration plan", "implementation plan", "design doc", "refactor plan", "propose a design", "feature plan", "architecture plan". Do NOT use for small fixes, one-line changes, or pure research questions.
Review project documentation for accuracy, completeness, clarity, and structure. Verifies file references, tests commands, validates links. Use for documentation audits, README reviews, or staleness checks. Triggers on "review docs", "audit documentation", "check README".
Review LLM-generated plans for autonomous agent executability using 8-dimension rubric. Triggers: "review plan", "compare plans", "plan quality", "meta-review".
Create production-ready v3.0 Cursor rule files by orchestrating template generation, schema validation, and RULES_INDEX.md indexing. Triggers on keywords like "create rule", "add rule", "new rule", "generate rule". Supports Python, Snowflake, JavaScript, Shell, Docker, Golang domains (000-999 range).
Determines which rule files to load for a given user request by matching file extensions, directory paths, and keywords against RULES_INDEX.md. Handles foundation loading, domain matching, activity matching, dependency resolution, and token budget management. Use when loading rules, selecting rules for a task, resolving rule dependencies, or managing token budgets during rule loading.
Execute agent-centric rule reviews (FULL/FOCUSED/STALENESS modes) using 6-dimension rubric and write results to reviews/rule-reviews/ with no-overwrite safety. Use when reviewing rule files, auditing rule quality, checking rule staleness, validating rule compliance, or analyzing agent executability.
Measures skill execution time and tracks performance. Use when timing a skill, measuring duration, comparing performance across models, analyzing execution speed, or detecting agent shortcuts.