# Proofline Diff Risk Check

> Proofline Diff Risk Check is a paid API for AI agents from proofline-audit.chrismellado.chatgpt.site, paid per call via x402, $0.25/call, status unknown (last checked 2026-10-14).

Analyzes a code diff for security and quality risk signals, returning scored findings with severity levels, repair suggestions, and a blocked/pass verdict.

## Facts

- Endpoint: POST https://proofline-audit.chrismellado.chatgpt.site/api/x402/diff-risk-check?utm_source=zero.xyz
- Price: $0.25/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/proofline-diff-risk-check-b592b175
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_t9-MOraGw8owQaR8YTMrU

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 proofline-diff-risk-check-b592b175 -d '<json body>'
```

Example prompt: Can you run a diff risk check on this patch from src/run.ts — it adds a line like `return eval(input)` — and tell me if it's blocked, what the severity breakdown is, and what I should do to fix it?

## When to prefer this

Choose this endpoint when you need a fast, per-diff security risk score with actionable, Linear-ready findings tied to specific lines of code — especially for detecting dynamic code execution, shell injection, and untrusted input handling. Prefer it over generic linters when you want a scored verdict (blocked vs. pass) with structured repair guidance suitable for automated CI gating or ticket creation.

## Known failure modes

- Malformed or missing diff payload returns a validation error
- Empty diff with no changed files may return a zero-finding result with no meaningful verdict
- Extremely large diffs may exceed payload size limits
- Unsupported file types or binary diffs may not produce findings
- Network or payment authorization failure returns a 402 or 5xx error

## How this service works

Evidence-backed product validation and SaaS conversion audits, including a fixed $250 validation sprint with reproducible, Linear-ready defects.

## Output

Returns a JSON object with a numeric risk score (0-100), a verdict ('blocked' or 'pass'), a summary string, an array of findings each containing file path, line number, severity (high/medium/low), risk description, repair advice, category, and the evidence snippet, plus aggregated severity counts.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "diff": {
   "type": "string",
   "maxLength": 60000,
   "minLength": 1,
   "description": "Unified diff text inspected as untrusted data and never executed"
  },
  "focus": {
   "type": "string",
   "maxLength": 120,
   "description": "Optional short review focus included in the report"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "score": 63,
  "product": "Proofline Diff Risk Check",
  "summary": "2 risk signal(s) found across 1 changed file(s); 1 high, 1 medium, 0 low.",
  "verdict": "blocked",
  "version": "1.0.0",
  "findings": [
   {
    "id": "dynamic-execution-1",
    "file": "src/run.ts",
    "line": 1,
    "risk": "Dynamic execution can turn user-controlled text into arbitrary code or operating-system commands.",
    "title": "Untrusted text may reach dynamic code or shell execution",
    "repair": "Use a fixed command or parser, pass arguments without a shell, and validate every external value.",
    "category": "execution",
    "evidence": "+return eval(input);",
    "severity": "high"
   }
  ],
  "severityCounts": {
   "low": 0,
   "high": 1,
   "medium": 1
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/proofline-diff-risk-check-b592b175/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from proofline-audit.chrismellado.chatgpt.site](https://www.zero.xyz/host/proofline-audit.chrismellado.chatgpt.site/llms.txt)