verification-before-completion-extras

What problem does it solve?

This Skill solves the common problem of accepting unverified completion, handoff, and derivation claims that lead to defective deliverables, rework, and incorrect downstream outputs. Without falsifiable verification, teams often trust narrative status updates instead of checking actual output artifacts, missing missing work, stale data, or broken guards.

Core Features & Use Cases

• Artifact-Anchored Handoff Verification: Validate delegated or parallel work against named output artifacts instead of trusting handoff narratives, ensuring no missing or partial work is accepted as complete.
• Dependency-Graph Freshness Checks: Determine if outputs are current by analyzing read/write dependencies between artifacts instead of relying on unreliable file timestamps.
• Detector Efficacy Validation: Prove that guards, regression tests, and other checks actually catch intended defects by running them against known broken states, not just clean inputs.
• Use Case: When a subagent hands off a financial model report claiming it is complete, use this Skill to verify the actual output file exists, contains all required candidate data, and that any attached guard tests fail on known defective model states before accepting the deliverable.

Quick Start

Use the verification-before-completion-extras skill to validate that a delegated team's completed deliverable matches the required output artifact and includes all expected data points before accepting the handoff.