# Eval Case Generator for AI Agents

> Eval Case Generator for AI Agents is a paid API for AI agents from relay402.georgespring.workers.dev, paid per call via x402, $0.02/call, status unknown (last checked 2026-09-14).

Generates happy-path, ambiguous-input, and adversarial-instruction eval test cases from a natural-language agent goal

## Facts

- Endpoint: GET https://relay402.georgespring.workers.dev/api/agent-eval-case-generate
- Price: $0.02/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/eval-case-generator-for-ai-agents-a7f3adc9
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_T4J2mqMGfpnaZEx4TRSq2

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability eval-case-generator-for-ai-agents-a7f3adc9
```

Example prompt: Generate a set of eval test cases — including a happy path, an ambiguous input scenario, and an adversarial instruction attempt — for an agent whose goal is 'book a restaurant reservation on behalf of the user'.

## When to prefer this

Choose this endpoint when you need deterministic, structured eval cases covering three specific risk axes (happy path, ambiguity, adversarial) without manually crafting prompts. It's especially useful for agent developers building automated test pipelines, red-teaming workflows, or regression suites. Prefer it over general-purpose LLM prompting when you need reproducible, categorized eval coverage from a single goal description.

## Known failure modes

- Goal string too short (< 3 chars) — returns validation error
- Goal string too long (> 4000 chars) — returns schema error
- Context string too long (> 16000 chars) — rejected
- Payment not provided or insufficient USDC — 402 Payment Required response
- Malformed request missing required 'goal' query param — 400 error
- Network timeout on the Cloudflare Workers edge — 5xx response

## How this service works

deterministic eval-case generator for agents: creates happy-path, ambiguous-input and adversarial-instruction cases from a goal.

## Output

Returns a structured set of three eval cases derived from the provided goal: (1) a happy-path case with a clear, well-formed input and expected successful output, (2) an ambiguous-input case designed to stress-test the agent's handling of unclear or underspecified requests, and (3) an adversarial-instruction case that attempts to manipulate or jailbreak the agent. Each case includes the scenario description and expected agent behavior.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "required": [
  "input"
 ],
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "type",
    "method"
   ],
   "properties": {
    "type": {
     "type": "string",
     "const": "http"
    },
    "method": {
     "enum": [
      "GET"
     ],
     "type": "string"
    },
    "queryParams": {
     "type": "object",
     "$schema": "https://json-schema.org/draft/2020-12/schema",
     "required": [
      "goal"
     ],
     "properties": {
      "goal": {
       "type": "string",
       "maxLength": 4000,
       "minLength": 3
      },
      "context": {
       "type": "string",
       "maxLength": 16000
      }
     },
     "additionalProperties": false
    }
   },
   "additionalProperties": false
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/eval-case-generator-for-ai-agents-a7f3adc9/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from relay402.georgespring.workers.dev](https://www.zero.xyz/host/relay402.georgespring.workers.dev/llms.txt)