verification-before-completion

Quick Review

Excellent skill addressing a critical AI behavior pattern - premature completion claims without verification. The description clearly conveys when to invoke this (before any completion/success claims). Structure is logical with clear sections, tables, and actionable patterns. Task knowledge is comprehensive with the 5-step Gate Function, concrete examples of claims vs requirements, and anti-patterns. The red flags and rationalization prevention sections provide strong behavioral guidance. Novelty is solid - while verification itself isn't novel, codifying this as a mandatory pre-completion gate that addresses AI tendency toward unverified assertions is valuable and would reduce wasted tokens on false completion cycles. Minor opportunity: could explicitly list exact verification commands for common scenarios (pytest, npm test, cargo build, etc.) though the principles clearly generalize. Very well-structured for AI consumption with clear rules, examples, and the critical 'spirit over letter' emphasis.