Parent
Subtask of qa-agent-workflows — Phase 4.
Problem
Even with structural and consistency checks, factual accuracy remains unverified. Claims may not match cited sources, dates may be wrong, quotes misattributed, and the KB may contain internal contradictions.
Solution
A research agent with web search capability that verifies specific claims against cited sources, checks historical accuracy, and detects cross-KB contradictions. Produces confidence-scored factual assessments with source chains.
Acceptance Criteria
Research agent with web search for claim verification
Cross-KB contradiction detection
Source chain verification (do cited sources actually support the claims?)
Confidence-scored factual assessments with source provenance
Checks: claim-source alignment, date accuracy, quote attribution, causal defensibility, statistic verifiability
CLI command: `pyrite qa verify [--kb ] [--entry ]`
Plugin architecture: domain-agnostic core with pluggable evaluation rubrics (legal, scientific, investigative)Dependencies
Phase 3 (qa-phase-3-tier-2-llm-assisted-consistency-checks) — builds on Tier 2 infrastructure
Web search capabilityFiles Likely Affected
Modified: `pyrite/services/qa_service.py` (Tier 3 verification logic)
New: verification agent module with web search integration
Modified: `pyrite/server/mcp_server.py` (qa verify tool)
Modified: `pyrite/cli/__init__.py` (qa verify command)