Back to marketplace
RAGPersonalHorizontal
RAG Faithfulness
Catch hallucinations before your users do.
Retrieval-augmented generation tasks with human-labeled ground truth. Measures answer faithfulness, citation accuracy, and abstention rate on unanswerable queries. Catches hallucinations before users do.
What's in the box.
Human-labeled gold answers with span-level citations
Includes deliberately unanswerable queries to test abstention
Separate scores for faithfulness, relevance, and completeness
InstallThree commands.
Three commands.
Then receipts.
Install the Pistachio CLI, add the harness as a Claude Code MCP tool, run it against your agent, and get a signed pass/fail report you can drop into a PR or sales deck.
CLI (Claude Code)
zsh
# 1. Install Pistachio CLI
npx @pistachio/cli login
# 2. Add the harness as a Claude Code MCP tool
pistachio harness add rag-faithfulness
# 3. Run it from inside Claude Code — get a signed report
pistachio harness run rag-faithfulnessSDK (Node)
typescript
import { Pistachio } from "@pistachio/sdk";
const pistachio = new Pistachio({ apiKey: process.env.PISTACHIO_KEY });
const run = await pistachio.harnesses.run("rag-faithfulness", {
endpoint: "https://your-agent.example.com/v1/messages",
});
console.log(run.passRate);
console.log(run.signedReportUrl);MCP-native
Lives inside Claude Code.
Signed reports
Ed25519 attestation.
Deterministic
Same input, same score.