LLM Output Validation

Instructions don't bind models. Checks in code do.

Schema validation with first-attempt and after-retry rates, invariant checks for judges and revisers, targeted repairs for catalogued failures, multi-stage extraction, and acceptance thresholds before writes.

Why it's different

A judge model broke its one rule in 35 of 49 reviews
First-attempt and after-retry rates logged separately
Python patterns included

Try asking

“structured output validation”“LLM as a judge”“extraction pipelines”

Details

Works with: Claude.ai, Claude Desktop, Claude Code, Codex and other SKILL.md agents
Needs: None
Pack: LLM Evaluation & Fact-Checking (€39)
Also covers: llm output validation, llm as a judge, structured output json

Same pack

Also in LLM Evaluation & Fact-Checking

LLM Evaluation Integrity

Eval rules that produced 55+ auditable runs and zero retracted numbers.

No setup neededView →

Fact-Check Before Sending

Catch the wrong number, name or organisation before it reaches an inbox.

No setup neededView →