jevrecipes

Recipe catalog / instruction-conflict

Check instructions for conflict

Decide whether two instructions can both be followed under the supplied circumstances.

You need to detect conflicting instructions before carrying out a task.

Explore this recipe interactively ยท Source and implementation guide

Use instruction-conflict in TypeScript

Install with npm install jev-recipes. Requires Node.js 22.9 or newer and ES modules. Set TYPESAFE_API_KEY in your server environment for live calls, which send input to TypeSafe and use API quota. See the installation guide.

import { instructionConflict } from 'jev-recipes/instruction-conflict';

const result = await instructionConflict({
  "firstInstruction": "Send the report as a PDF attachment.",
  "secondInstruction": "Send the report only as plain text in the email body; do not attach any files.",
  "context": "Both instructions apply to the same outgoing report email."
});
console.log(result);

Input contract

FieldTypeNeeded
firstInstructionstringRequired
secondInstructionstringRequired
contextstringOptional
minConfidencenumberOptional
Full input and result schemas
{
  "input": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "type": "object",
    "properties": {
      "firstInstruction": {
        "type": "string",
        "description": "The first instruction to compare."
      },
      "secondInstruction": {
        "type": "string",
        "description": "The second instruction to compare."
      },
      "context": {
        "type": "string",
        "description": "The task and circumstances in which the instructions may apply."
      },
      "minConfidence": {
        "type": "number",
        "minimum": 0,
        "maximum": 1
      }
    },
    "required": [
      "firstInstruction",
      "secondInstruction"
    ]
  },
  "result": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "type": "object",
    "properties": {
      "model": {
        "type": "string"
      },
      "usage": {
        "type": "object",
        "properties": {
          "input_tokens": {
            "type": "integer",
            "minimum": 0,
            "maximum": 9007199254740991
          },
          "output_tokens": {
            "type": "integer",
            "minimum": 0,
            "maximum": 9007199254740991
          }
        },
        "required": [
          "input_tokens",
          "output_tokens"
        ],
        "additionalProperties": false
      },
      "status": {
        "type": "string",
        "enum": [
          "ready",
          "review"
        ]
      },
      "verdict": {
        "type": "string",
        "enum": [
          "compatible",
          "conflicting",
          "different_scope",
          "unclear"
        ]
      },
      "confidence": {
        "type": "number",
        "minimum": 0,
        "maximum": 1
      },
      "probabilities": {
        "type": "object",
        "propertyNames": {
          "type": "string",
          "enum": [
            "compatible",
            "conflicting",
            "different_scope",
            "unclear"
          ]
        },
        "additionalProperties": {
          "type": "number",
          "minimum": 0,
          "maximum": 1
        },
        "required": [
          "compatible",
          "conflicting",
          "different_scope",
          "unclear"
        ]
      }
    },
    "required": [
      "model",
      "usage",
      "status",
      "verdict",
      "confidence",
      "probabilities"
    ],
    "additionalProperties": false
  }
}

Saved example result

This hand-authored response demonstrates the contract. It is not a model accuracy measurement. Run it without an API key: npx jev-recipes demo instruction-conflict.

{
  "model": "demo-fixture",
  "usage": {
    "input_tokens": 0,
    "output_tokens": 0
  },
  "status": "ready",
  "verdict": "conflicting",
  "confidence": 0.96,
  "probabilities": {
    "compatible": 0.013333333333333334,
    "conflicting": 0.96,
    "different_scope": 0.013333333333333334,
    "unclear": 0.013333333333333334
  }
}

Evaluation evidence

Fixture only

No verified live accuracy measurement is available. Evaluate representative cases before using this decision in your workflow.

Use the evaluation guide to measure this decision on your own labeled cases.

Limitations

Related recipes