---
name: agent-explainability-specialist
description: Use for designing implementing and enforcing complete explainability transparency and interpretability in AI agents multi-agent systems skill creation decision pipelines and outputs. Triggers include explainability XAI agent reasoning trace causal explanation decision audit trail why-agent-did-this skill explainability complete explainability skill-creator transparency or any request for making agent behavior fully understandable verifiable and auditable by humans or other agents. Essential for high-stakes agent economies compliance certification and trust in 2026.
metadata:
  type: specialist
  version: "1.0"
  category: governance-alignment
  income-priority: high
  integrates-with: agent-auditor-compliance-specialist, ethical-agent-behavior-designer-alignment-specialist, agent-evaluation-red-teaming-certification-specialist, skill-creator, empireforge-meta-governor-operator-god-mode-dashboard, recursive-skill-evolution-orchestrator, agent-regulatory-compliance-specialist, agent-super-alignment-specialist, agent-super-nvidia-specialist
---

# Agent Explainability Specialist

## Overview
Specialize agents and skill-creation workflows so every decision has complete machine-readable + human-readable explainability (reasoning traces, causal chains, counterfactuals, confidence scores, audit-ready narratives).

## When to Activate
Any skill creation, high-stakes decisions, audits, certification, MCP/A2A/Cloudflare Agent Ready work.

## Mandatory Components
1. Reasoning Trace JSON (trace_id, steps with confidence, alternatives, final_decision)
2. Causal Explanation + Counterfactual
3. Human Narrative Layer
4. Protocol Attachments (X-Agent-Explainability, MCP, Content-Signal)
5. Certification & Audit Hooks

## skill-creator Integration
Force Explainability Contract + metadata.explainability into every skill. Validate 7/7 with validate-explainability.sh.

## Scripts and Assets
- scripts/generate-trace.sh
- scripts/validate-explainability.sh
- references/xai-standards-2026.md
- references/counterfactual-templates.md
- assets/trace-schema.json

## Related Skills
Always compose with agent-regulatory-compliance-specialist, agent-super-alignment-specialist, agent-super-nvidia-specialist.

This skill makes every other skill and every agent decision trustworthy, certifiable, and monetizable.