prompt-injection-detector
Prompt Injection Detector is a stateless security microservice built for the agent economy. Before your agent acts on any input — whether from a user, an email, a document, or another agent — run it through this scanner first. Detects: jailbreak attempts, instruction overrides, role-play manipulation, goal hijacking (OWASP ASI01), tool misuse patterns (OWASP ASI02), hidden instructions embedded in data, and indirect injection via external content. Returns a confidence score, severity rating, specific patterns detected, and a clear recommendation — block, warn, or pass. Pay per scan via x402 protocol. USDC on Base. No API keys. No subscriptions. No human in the loop. Agents call it directly. MCP tool name: scan_for_injection
Composite of vulnerability cleanliness, spec conformance, provenance, stability, and usage signals — scanned and weighted by Cognium. Human and agent signals are tracked separately.
View full trust & usage report →Metadata
- Version
- 1.0.0
- Skill type
- atomic
- Execution layer
- mcp-remote
- Category
- communication
- Source
- Smithery
- Author type
- human
- Updated
- 2026-05-12
Use via MCP
Resolve prompt-injection-detector from your agent
Streamable HTTP transport at https://api.skillsregistry.net/mcp. No auth for read tools. Discovery: .well-known/mcp.json.
One command in your shell — Claude Code wires it up and verifies the connection. Run /mcp in any session to confirm.
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp--scope user for --scope project to commit it to .mcp.json.