P2 DR Production Flow

Generates diagnostic and rationale sections through multi-agent threaded convergence auditing.

Updated Apr 11, 2026
One-click install
npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill p2-dr-production-flow-gozonerd
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: P2 DR Production Flow
Source: https://github.com/gozonerd/shadow-ai-assessment/tree/main/.claude/skills/dmis/pipeline/P2_DR_Production_Flow
Command: npx skills add https://github.com/gozonerd/shadow-ai-assessment --skill p2-dr-production-flow-gozonerd

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Producing evidence-backed diagnostic and rationale sections requires consistent, verifiable reasoning across many generated candidates, and single-pass generation often yields unverified claims, contradictions, or incomplete coverage. ## Core Features & Use Cases - Multi-Phase Pipeline: Runs DR prompt creation, 25 parallel thread generations, an 11-layer convergence audit, synthesis, and post-synthesis checks in a defined sequence. - Convergence Quality Control: Enforces a Krippendorff alpha threshold of 0.70 across 11 audit layers covering compliance, factual agreement, contradiction detection, and reference integrity. - Quality Gate Enforcement: Blocks sections that fail post-synthesis checks, fall below the 15 KB minimum size, or contain untraceable evidence citations. - Use Case: After a P1 analysis section identifies diagnostic claims, run this pipeline to produce a verified dr_synthesis section where every claim traces back to the evidence library and survives layered adversarial audits. ## Quick Start Run the P2 DR production flow on the completed P1 section output and evidence library to generate the audited diagnostic and rationale synthesis.

Frequently Asked Questions about P2 DR Production Flow

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a diagnostic rationale section from P1 analysis output?

Feed the P1 section output, evidence library, and diagnostic requirements into the P2 pipeline. It creates a tailored DR prompt, generates 25 candidate threads in parallel, audits them across 11 layers, and synthesizes the winning thread into a dr_synthesis section.

What is a convergence audit in multi-thread generation?

A convergence audit runs 11 sequential validation layers over 25 candidate threads, checking compliance, factual agreement, contradictions, and reference integrity. Agreement is measured with Krippendorff alpha, requiring a threshold of 0.70 or higher to proceed to synthesis.

What happens when a convergence audit layer fails?

A failed layer triggers one critique-rewrite retry cycle, and a second failure on the same layer escalates to the convergence judge. The judge decides between remediation and halting, with quality gate failures escalated to a human decision.

Why does the quality gate reject outputs under 15 KB?

The pipeline enforces a dr_synthesis minimum of 15360 bytes because diagnostic sections are evidence-rich, and undersized outputs indicate incomplete coverage. Any final output below 15 KB fails the quality gate automatically.

Can P2 output remain valid if P1 is revised afterward?

No. P2 is tightly coupled to P1, and all diagnostic claims must trace back to P1 observations. If P1 output is revised after P2 completes, the entire P2 pipeline must be re-run with the updated evidence context.