What problem does it solve? AI capability evolves faster than a model's knowledge cutoff, so feasibility judgments and time estimates based on training-era assumptions are systematically wrong—usually far too pessimistic. This Skill forces verification against current capability before scheduling, estimating, or judging AI-dependent work. ## Core Features & Use Cases - Capability Verification: Searches shared memory and the web for recent model releases, coding-agent benchmarks, and automation tooling before making feasibility calls. - Calibration Discipline: Assumes current capability is at least an order of magnitude beyond cutoff-era expectations and reports verified facts rather than assumptions. - Estimate Guardrails: Avoids promising durations; gives time estimates only when explicitly requested, labeled with uncertainty ranges. - Use Case: Before telling a stakeholder that an AI agent cannot reliably automate a workflow, invoke this Skill to check the latest benchmarks and model releases, then state what was verified. ## Quick Start Before estimating this AI automation project, check current AI agent capability and tell me what is actually feasible today.
