Gsep Mcp
A security layer for AI agents that runs through MCP instead of adding more tools. Exposes 10 security-focused operations including input scanning (C3 firewall with 53 prompt injection patterns), response integrity checks (C4 immune system), action blocking (C5 guard with 80+ destructive patterns), and self-evolving prompt management across six chromosome layers. Works over stdio for Claude Desktop, Cursor, Windsurf, and other IDEs, or as an HTTP server with an OpenAI-compatible gateway for n8n, Make, and backend integrations. Supports Claude, GPT-4, Gemini, Ollama, and Perplexity. Useful when you need protection from injection attacks and unsafe operations without modifying your agent's core logic or workflow.
- tools
- 10
- auth
- Secret needed
- transport
- STDIO
- env vars
- 4
- last updated
- May 5, 2026
Tools
10 toolsPublic tool metadata for what this MCP can expose to an agent.
gsep_chatFull pipeline — C3 → evolved LLM → C4 → C5 → fitness → evolution.
Full pipeline — C3 → evolved LLM → C4 → C5 → fitness → evolution.
No parameter schema in public metadata yet.
gsep_scan_inputC3 Content Firewall — scan any text before sending to your LLM.
C3 Content Firewall — scan any text before sending to your LLM.
No parameter schema in public metadata yet.
gsep_scan_outputC4 Behavioral Immune System — verify your LLM's response wasn't manipulated.
C4 Behavioral Immune System — verify your LLM's response wasn't manipulated.
No parameter schema in public metadata yet.
gsep_scan_actionsC5 Action Firewall — catch dangerous commands before they run.
C5 Action Firewall — catch dangerous commands before they run.
No parameter schema in public metadata yet.
gsep_before_llmMiddleware pre-hook — sanitize input and assemble the protected prompt before an external agent calls its own LLM.
Middleware pre-hook — sanitize input and assemble the protected prompt before an external agent calls its own LLM.
No parameter schema in public metadata yet.
gsep_after_llmMiddleware post-hook — verify an external LLM response before showing it to users or tools.
Middleware post-hook — verify an external LLM response before showing it to users or tools.
No parameter schema in public metadata yet.
gsep_before_toolTool execution pre-hook — block dangerous shell, database, filesystem, or API actions before execution.
Tool execution pre-hook — block dangerous shell, database, filesystem, or API actions before execution.
No parameter schema in public metadata yet.
gsep_after_toolTool result post-hook — treat tool output as untrusted external content before reinjecting it into the agent.
Tool result post-hook — treat tool output as untrusted external content before reinjecting it into the agent.
No parameter schema in public metadata yet.
gsep_get_statusGenome health, fitness scores, drift detection, evolution generation.
Genome health, fitness scores, drift detection, evolution generation.
No parameter schema in public metadata yet.
gsep_record_feedbackSignal satisfaction/dissatisfaction to drive evolution.
Signal satisfaction/dissatisfaction to drive evolution.
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageConfiguration
ANTHROPIC_API_KEYsecretAnthropic API key (use this OR OPENAI_API_KEY OR OLLAMA_HOST)
OPENAI_API_KEYsecretOpenAI API key (use this OR ANTHROPIC_API_KEY OR OLLAMA_HOST)
OLLAMA_HOSTOllama server URL for local models (e.g. http://localhost:11434)
GSEP_PRESETIntelligence preset: minimal, standard, conscious, or full (default: full)
Readme
GSEP-MCP — AI Agent Security via Model Context Protocol
The only MCP server that protects your AI agent instead of just extending it.
"me encanta saber que no borrará nada de mi pc" — First GSEP user, unprompted
At a Glance
Details
- Repository
- github.com/gsepcore/gsep-mcp
- Categories
- AI & LLM Tools, Security & Pentesting
- Package
@gsep/mcp- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active
Related AI & LLM Tools MCP servers
View all
DrinkedIn — licensed bar dataio.github.haim-barad/drinkedinLicensed bar data from public registers, x402. Agent-bar tools deprecated. No real alcohol.0stars
Foresea Forecastingpareelamre/analyzing-llm-rationaleForecast future events and scan prediction-market edges.0stars
Pentagonal Mcppentagonal-ai/pentagonalAI smart contract forge — 8-agent security audits, generation, and compilation across 8 chains0stars
AgentTwinagenttwin/agenttwin-agent-toolsTest AI agent workflows against service twins with injected faults. Runs locally; no source upload.0stars
Mcp Serverarkheiaai/arkheia-mcpDetect fabrication and hallucination in any LLM output. 30+ models supported.0stars
Attestattestagents/install-attest-agentScan agent payment endpoints (x402, AP2, L402, MPP, 402) for a trust grade A-F before paying.0stars