LLM Consistency Testing - QA Skill Library

Why this Skill

It turns this Skill's method into a quality input that can be executed, reviewed, and reused.

Use this skill when you need evidence-bounded analysis, design, or validation preparation for repeat inputs, version/model/prompt factors, invariants, variance evidence, and comparison boundaries.
Keep the analysis focused on repeat inputs, version/model/prompt factors, invariants, variance evidence, and comparison boundaries; do not replace business owners or Human risk acceptance, exception approval, or safety decisions.
Never invent system behavior, fields, model outputs, data, thresholds, root causes, execution records, or pass claims.
Static design, plans, file presence, or a dry run retain their evidence state and cannot become proof of real execution.
Listing checks without triggers, expected concerns, owner roles, close conditions, and evidence.

When to use

Use this Skill

Use this skill when you need evidence-bounded analysis, design, or validation preparation for repeat inputs, version/model/prompt factors, invariants, variance evidence, and comparison boundaries.
Use it to review an Agent, RAG, or LLM plan, result, or evidence set and produce actionable improvements.
Use it when context is incomplete but a bounded first pass with assumptions, gaps, and human-decision boundaries is still useful.

Common pitfalls

Listing checks without triggers, expected concerns, owner roles, close conditions, and evidence.
Treating adjacent tests or model tools as a complete LLM consistency judgment.
Using unexplained numbers for false precision or writing correlation as causation.
Refusing incomplete input, or pretending that incomplete evidence is conclusive.

Input

Minimum Input

The current task scope, objective, and subject under review.

Recommended Input

Project goal
Test scope
Constraints

Optional Context

Relevant code or configuration
Historical results
Logs and metrics

Output

The output follows this Skill's method and makes facts, assumptions, risks, and next steps explicit.

Judge the output value before installing

01Read and follow prompts/llm-consistency-testing.md, including its input audit, domain coverage, and output order.
02Extract scope, environment, version, time window, constraints, success criteria, and available evidence, with attention to repeat input, version factor, model factor, invariant, variance evidence.
03Separate confirmed facts, evidence-backed inferences, candidate recommendations, and Human decisions before ranking by risk and evidence strength.
04Turn high-risk items into preconditions, steps, expected behavior or decision criteria, required evidence, and a validation method.

View full output structure

05When input is incomplete, deliver a bounded first pass, state unsupported conclusions, and never present static material as execution evidence.

How It Works

01Read and follow prompts/llm-consistency-testing.md, including its input audit, domain coverage, and output order.
02Extract scope, environment, version, time window, constraints, success criteria, and available evidence, with attention to repeat input, version factor, model factor, invariant, variance evidence.
03Separate confirmed facts, evidence-backed inferences, candidate recommendations, and Human decisions before ranking by risk and evidence strength.
04Turn high-risk items into preconditions, steps, expected behavior or decision criteria, required evidence, and a validation method.
05When input is incomplete, deliver a bounded first pass, state unsupported conclusions, and never present static material as execution evidence.

Install & Quick Start

Install command / SHELL

npx skills add \
  https://github.com/naodeng/awesome-qa-skills/tree/main/skills/en/testing-types/llm-consistency-testing
  -g

llm-consistency-testing.prompt

@skill llm-consistency-testing

Using the current project context, produce an actionable result following this Skill.

Additional context:
[Paste project context or requirement]

---
name: llm-consistency-testing
description: Use this skill when you need evidence-bounded repeat inputs, version/model/prompt factors, invariants, variance evidence, and comparison boundaries; triggers include LLM 一致性 and LLM consistency.
---

# LLM Consistency Testing

## When to Use

- Use this skill when you need evidence-bounded analysis, design, or validation preparation for repeat inputs, version/model/prompt factors, invariants, variance evidence, and comparison boundaries.
- Use it to review an Agent, RAG, or LLM plan, result, or evidence set and produce actionable improvements.
- Use it when context is incomplete but a bounded first pass with assumptions, gaps, and human-decision boundaries is still useful.

## Output Format Options

- Default to Markdown organized by domain risk, evidence state, priority, and boundary.
- When the user requests tables, CSV, JSON, or ticket fields, preserve the same finding fields, evidence, and decision boundaries.
- Before machine consumption, confirm the schema, enums, required fields, and evidence sources.

## How to Use

1. Read and follow `prompts/llm-consistency-testing.md`, including its input audit, domain coverage, and output order.
2. Extract scope, environment, version, time window, constraints, success criteria, and available evidence, with attention to repeat input, version factor, model factor, invariant, variance evidence.
3. Separate confirmed facts, evidence-backed inferences, candidate recommendations, and Human decisions before ranking by risk and evidence strength.
4. Turn high-risk items into preconditions, steps, expected behavior or decision criteria, required evidence, and a validation method.
5. When input is incomplete, deliver a bounded first pass, state unsupported conclusions, and never present static material as execution evidence.

## Reference Files

- Always read `prompts/llm-consistency-testing.md`; it is the complete execution specification for this skill.
- For evaluation, read `evals/eval.yaml` and the matching cases under `evals/cases/`.
- Load `references/`, `examples/`, `scripts/`, or `output-formats.md` only when those directories exist and the task needs them.

## Core Constraints

- Keep the analysis focused on repeat inputs, version/model/prompt factors, invariants, variance evidence, and comparison boundaries; do not replace business owners or Human risk acceptance, exception approval, or safety decisions.
- Never invent system behavior, fields, model outputs, data, thresholds, root causes, execution records, or pass claims.
- Static design, plans, file presence, or a dry run retain their evidence state and cannot become proof of real execution.
- When evidence is insufficient, use pending confirmation, blocked, unassessed, or NOT_SCORED and give the smallest validation method.
- For user data, production, or safety work, use least privilege, masked data, mocks, dry runs, or isolation.

## Delivery Checklist

- [ ] Covered repeat input, version factor, model factor, invariant, variance evidence, with source, evidence state, and validation method for each.
- [ ] Separated facts, inferences, candidate recommendations, gaps, and Human decisions.
- [ ] Gave high-risk items P0/P1/P2/P3 or an equivalent priority, owner role, and close condition.
- [ ] Did not turn plans, static checks, or dry runs into test execution, all-passed, or safety-approved claims.
- [ ] Stated residual risk, stop/escalation conditions, and next actions.

## Common Pitfalls

- Listing checks without triggers, expected concerns, owner roles, close conditions, and evidence.
- Treating adjacent tests or model tools as a complete LLM consistency judgment.
- Using unexplained numbers for false precision or writing correlation as causation.
- Refusing incomplete input, or pretending that incomplete evidence is conclusive.

## Best Practices

- Start with paths most likely to cause user harm, business loss, or decision blockage.
- Use the smallest verifiable experiment to reduce uncertainty and record conditions, versions, sources, and evidence.
- Make the Skill independently installable, executable, and reviewable by another engineer.