# RelayShield Agent Bait Scan

> RelayShield Agent Bait Scan is a paid API for AI agents from api.relayshield.net, paid per call via x402, $0.5/call, status unknown (last checked 2026-10-27).

Scans a GitHub repository's agent instruction files for prompt injection and bait instructions that could manipulate AI agents

## Facts

- Endpoint: POST https://api.relayshield.net/v1/payg/agent-bait-scan?utm_source=zero.xyz
- Price: $0.5/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-27
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/relayshield-agent-bait-scan-91e5b739
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_7-9mAdQM1xqxYPmK8pCFn

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 relayshield-agent-bait-scan-91e5b739 -d '<json body>'
```

Example prompt: Before I connect my Cursor agent to that new MCP server, can you scan the GitHub repo at https://github.com/someorg/some-mcp-server for any bait instructions or hidden prompt injection attacks in its README, AGENTS.md, or other agent config files?

## When to prefer this

Choose this endpoint when you need to vet the safety of a GitHub repository's AI agent configuration files — specifically README, AGENTS.md, CLAUDE.md, .cursorrules, and MCP manifests — before allowing an AI agent to consume or act on them. It is purpose-built for detecting prompt injection and bait instructions targeting agentic AI systems, making it the right choice over generic malware scanners or URL checkers when the threat vector is adversarial agent instructions embedded in repositories.

## Known failure modes

- Repository not found or inaccessible — returns error if the GitHub URL is invalid or the repo is private
- No agent instruction files found in repository — scan completes but findings are empty
- Network timeout if repository is very large or GitHub rate limits apply
- Malformed repository identifier returns validation error
- False negatives possible for novel or obfuscated bait instruction patterns

## How this service works

Identity security, threat monitoring and intelligence for the agentic AI era. TAP verifier for agentic commerce, MCP Proxy Firewall for AI agents, and threat intelligence APIs backed by 725K indicators across 129 monitored marketplaces.

## Output

Returns a JSON object with an overall verdict (e.g. LOW, MEDIUM, HIGH) indicating the risk level of agent-bait or prompt injection found, plus a list of specific findings detailing any suspicious instructions discovered in the repository's agent instruction files.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "repository": {
   "type": "string",
   "description": "GitHub repository URL or owner/repo whose agent instruction files (README, AGENTS.md, CLAUDE.md, .cursorrules, MCP manifests and tool descriptions) to scan for bait instructions"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "ok": true,
  "data": {
   "verdict": "LOW",
   "findings": []
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/relayshield-agent-bait-scan-91e5b739/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.relayshield.net](https://www.zero.xyz/host/api.relayshield.net/llms.txt)