Skip to main content
Glama
genpark-agent-exfiltration-canary-leak-detector-skill

genpark-agent-exfiltration-canary-leak-detector-skill

Official

genpark-agent-exfiltration-canary-leak-detector-skill

Organization Model Context Protocol Zero-Dependency Python Version Security Rating License

Production-hardened Agentic Cybersecurity & Red-Teaming skill engineered strictly with Python standard library. Provides real-time injection fuzzing, exfiltration canary detection, least-privilege tool sandboxing, and cryptographic audit proofs via the Model Context Protocol (MCP).


Architecture & Data Flow

flowchart TD
    Payload[Untrusted Prompt / RAG Context / Agent Action] --> Inspector[Telemetry Ingestion]
    Inspector --> Core["GenParkAgentExfiltrationCanaryLeakDetector"]
    Core --> Detection[Cryptographic / Heuristic Analysis Engine]
    Detection --> Mitigation[Least-Privilege & Boundary Enforcement]
    Mitigation --> MCPOutput[MCP Result / Quarantine Signal / Merkle Seal]

Related MCP server: Diogenes Agent Guardrails & Command Sandbox MCP

Quickstart & Usage

1. Direct Python Client Execution

python example_usage.py

2. Standalone MCP Server Launch

Run the native STDIO MCP server:

python mcp_server.py

License

This project is licensed under the MIT License - see the LICENSE file for details.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      Protects AI agents from prompt injection attacks, jailbreak attempts, and common web vulnerabilities by screening untrusted input through semantic LLM analysis and static pattern matching.
      68 npm
      2
      ISC
    • A
      license
      B
      quality
      B
      maintenance
      Enables pre-execution security checks and sandboxed execution for AI agent workflows, including prompt-injection detection, MCP rugpull and sampling abuse protection, untrusted tool-output sanitization, loop circuit breaking, SQL AST validation, PII redaction, and ephemeral Python execution.
      8
      2
      MIT
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enables red-team testing of AI agents against indirect prompt injection by fuzzing payloads with delimiter evasion and Unicode homoglyphs, detecting exfiltration canaries, enforcing least-privilege tool boundaries, and producing cryptographic audit proofs.
      7
      MIT