shamt-validation-loop

The Shamt validation loop protocol automates artifact validation by enforcing multi-round re-reads, dimension checks, and adversarial self-checks across specs, plans, guides, design docs, and code within your project.

Adversarial checks in validation loops perform rigorous self-assessments every round to proactively identify potential flaws in artifacts before progressing to sub-agent confirmations and final approval.

Enforcing consecutive clean rounds requires tracking the consecutive_clean metric throughout the validation loop and exiting only after a primary clean round is confirmed by two independent Haiku sub-agents.

Yes, the validation loop protocol applies to software code, implementation plans, specs, guides, and design docs within any project, performing tool-based verification and recording validation logs for consistent quality decisions.

The validation loop requires completing a primary clean round before initiating dual Haiku sub-agent confirmations, meaning the protocol cannot exit early and demands full multi-round processing to ensure rigorous quality assurance.