Hallucination Failure-Mode Audit Rubric
Audit any LLM feature against a 25-class hallucination taxonomy — run per-class detection probes, score every finding by severity and likelihood, and ship a prioritized per-feature risk register and remediation plan.
A 4-step agentic workflow pack for coding built to run with ChatGPT, Claude, and Gemini. Open the Markdown files, fill the variables, and paste into your model. Most buyers get a reviewable result in about 15 minutes.
- Audit any LLM feature against a 25-class hallucination taxonomy with a detection probe and mitigation for each class
- Scope the audit to the classes your feature can actually produce, so you probe what matters instead of running generic checks
- Grade captured outputs claim by claim — separating verified facts from unsupported assertions with cited evidence
- Score every finding on a calibrated severity x likelihood matrix and roll it up into a weighted per-feature risk register
- Prioritize mitigations by risk reduction versus effort, with owners, gates, and a re-audit trigger for each fix
- Standardize your team's hallucination review so results are consistent across engineers and across model versions
Prompt Customization Service — optional help adapting variables and output to your brand voice. Choose your tier at checkout (not tied to this prompt's price).
This pack is $10 on its own. Buying every pack separately costs $1076. The Lifetime Bundle is $149 one-time — you save $927 (86% off) and unlock every future pack free.
Get the Lifetime Bundle — $149Paste the license key from your receipt. It must match this prompt pack.
What ships with your purchase
Prompt files
Plain Markdown files with `{{variables}}` you fill in, ready to paste into ChatGPT, Claude, or Gemini. No setup, no tooling required.
Usage guide
Variable reference, model compatibility, examples, and customization tips so you can adapt the pack to your brand voice.
Lifetime updates
When we improve the pack, you get the new version automatically. Email support included with every purchase.
Models tested: ChatGPT, Claude, Gemini.
The workflow inside this pack
4 composable prompts you run in order — each one picks up where the last left off.
- Step 1
Scope and Criteria Builder
Paste a description of your LLM feature, its output surfaces, and the grounding sources it has, and get a scoped audit plan: which of the 25 hallucination classes apply, why, and the exact detection probe and pass criterion for each.
- Step 2 · optional
Evidence Interrogator
Feed in the audit plan plus a batch of captured outputs with their inputs and grounding context, and get a per-claim evidence table: each claim tagged with its hallucination class, a verified / unsupported / contradicted grade, and the exact source span that supports or refutes it.
- Step 3 · optional
Scored Findings Report
Submit the interrogated evidence and optional scoring weights, and get a scored risk register: each finding rated on severity and likelihood with a confidence level, rolled up into weighted per-class scores and an overall feature risk band.
- Step 4 · optional
Remediation Prioritizer
Feed in the scored risk register and your team's capacity, and get a prioritized remediation table: each fix scored on risk reduction and effort, mapped to the class-specific mitigation, with an owner, a gate, and the probe that re-verifies it.
Perpetual (lifetime) use license
Your one-time purchase includes an ongoing right to use this prompt pack with the AI tools and models you control for your own and your clients' work — not for resale or public redistribution of the files as a product.
We keep the copyright
The prompt files, guides, examples, and bundled assets stay our copyrighted works (or our licensors'). Payment grants the limited license in our Terms only — it does not transfer ownership.
Need help adapting this prompt to your team? Add Prompt Customization Service at checkout.