About

We build by decomposing strategy into briefs, briefs into tasks, and tasks into verified evidence. Each DEV session has acceptance criteria, a maximum task count, mandatory MCP logging, and a session report. Each ARC verify session checks acceptance criteria against reality. Each DEEP audit compares the new state to the previous baseline.

The result is an organization that improves through explicit feedback loops:

• Strategy → briefs
• Briefs → code + content + config
• Execution → evidence
• Evidence → verification
• Verification → audit
• Audit → strategy

This loop prevents drift. A feature that looks done but has no evidence fails verification. A fix that improves one metric while degrading another is caught by the DEEP audit. A decision that is not recorded in the wiki cannot inform the next agent.