What problem does it solve? After delivering a project or module, teams often lose the real numbers—actual hours, costs, and variance versus the original estimate—so future quotes stay based on gut feeling. This Skill captures those post-delivery actuals into a structured METRIC knowledge node, creating a calibration loop that improves estimation accuracy over time. ## Core Features & Use Cases - Structured Actuals Capture: Guides the user through five questions (actual hours, person-months, cost, variance, satisfaction) and auto-computes person-months and variance percentage against any existing EST estimate. - Standardized METRIC Node Generation: Writes a YAML-frontmatter Markdown file to knowledge/nodes/metrics/ with a deterministic ID format (METRIC-{TARGET}-{YYYYMMDD}), relations, and citation references. - Knowledge Graph Integration: Links the METRIC back to its target CASE/MOD/EST unit, runs citation validation and index rebuild scripts, and proposes follow-up DEC records for estimate calibration decisions. - Use Case: After a client CRM project is accepted, tell the assistant the project took 120 hours versus the 90-hour estimate; the Skill writes METRIC-CASE-RSUN-CRM-20260421.md, computes the +33% variance, and suggests adjusting the rate-card assumption for similar cases. ## Quick Start Tell the assistant to capture the metric for the just-finished case, providing the actual hours, cost, and a satisfaction score so it can write the METRIC node and compute variance against the estimate.
