Quality Assurance

The Quality Assurance skill acts as a critical quality gate within the PB-000 market research automated pipeline. It is designed to rigorously audit and evaluate markdown documentation produced by various agents throughout the research phases (Phase 1 through Phase 5). By comparing generated content against specific, pre-defined phase checklists, the skill ensures that all deliverables meet high standards of data integrity, specificity, and logical consistency. This agent is essential for teams aiming to maximize the success probability of Gate 1 reviews by providing objective, quantitative scoring and actionable improvement feedback.

• Performs automated parsing and analysis of project-specific markdown (.md) files.

• Validates required items vs. recommended items based on phase-specific checklist files (phase{X}_checklist.md).

• Calculates quantitative quality scores (out of 100) across three primary axes: completeness, specificity, and data-driven evidence.

• Generates structured, actionable improvement reports in markdown, identifying specific tasks for agents to perform before moving to the next phase.

• Enforces strict compliance with passing criteria (e.g., Phase 1-4 minimum 80 points, Phase 5 minimum 85 points).

• Targeted primarily at automated multi-agent market research workflows involving persona analysis, competitive intelligence, and positioning strategy.

• Requires existing phase-specific deliverables and quality standard files located in the project structure (pb000_deliverables/).

• Outputs status summaries (PASS, CONDITIONAL_PASS, FAIL) to guide the orchestration logic of the main agent.

• Highly effective for maintaining consistency when integrating data from APIs like Apify or search results from Gemini DeepSearch.

• Users should trigger this skill via natural language requests in Claude Code, specifying the target phase ID and the path to the artifacts requiring review.