Agent Output Guard
When you're running multi-agent systems and need to validate data at handoff boundaries, this server gives you five tools that work without LLM calls. You get verify_json_schema for structure checking, detect_hallucination_markers to scan for uncertainty phrases like "I think" or "probably", validate_data_freshness to check timestamps against max age thresholds, cross_reference_check to compare outputs from multiple agents, and output_consistency_score for overall reliability metrics. It's pure computation, so there's no API cost per validation. Reach for this when Agent A's output feeds Agent B and you need programmatic guards against format mismatches, stale data, or fabricated content before the receiving agent acts on it.
- tools
- 5
- transport
- STDIO
- last updated
- Apr 2, 2026
Tools
5 toolsPublic tool metadata for what this MCP can expose to an agent.
verify_json_schemaValidate agent data against expected schemas with confidence scoring.
Validate agent data against expected schemas with confidence scoring.
No parameter schema in public metadata yet.
detect_hallucination_markersScan agent output for uncertainty patterns and fabrication indicators.
Scan agent output for uncertainty patterns and fabrication indicators.
No parameter schema in public metadata yet.
validate_data_freshnessCheck if agent data is current and valid based on timestamps.
Check if agent data is current and valid based on timestamps.
No parameter schema in public metadata yet.
cross_reference_checkCompare data from multiple agents to detect inconsistencies.
Compare data from multiple agents to detect inconsistencies.
No parameter schema in public metadata yet.
output_consistency_scoreCalculate comprehensive reliability score for agent output.
Calculate comprehensive reliability score for agent output.
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 languageReadme
Agent Output Guard MCP Server 🛡️
The first MCP server designed specifically to solve coordination failures in multi-agent systems. Built by Agenson Horrowitz based on the MAST study showing 36.9% of multi-agent failures are coordination breakdowns.
🚨 The Multi-Agent Coordination Crisis
Details
- Repository
- github.com/agenson-tools/agent-output-guard-mcp
- Categories
- AI & LLM Tools, Data & Analytics
- Package
@agenson-horrowitz/agent-output-guard-mcp- Transport
- STDIO
- Registry
- MCP registry, active
More from agenson-tools
Related AI & LLM Tools MCP servers
View all
Vector Search APIbr0ski777/vector-search-x402In-memory vector search with TF-IDF and cosine similarity. x402 micropayment.0stars
Agent Toolbelt Mcpmarras0914/agent-toolbelt19 focused API tools for AI agents: contract review, prompt optimizer, context packing, and more.0stars
Agent Costcentermdfifty50-boop/agent-costcenter-mcpMulti-agent billing and cost allocation tracking0stars
Traced Llm Proxymikerawsonnz/gosce-agentsProxy Gemini (Vertex AI) completions wrapped in OpenTelemetry trace spans; returns the answer plus t0stars
agent-godmodemdvohra/agent-godmodeSandboxed workspace MCP: file ops, argv-only run_command, list_files; BYO LLM prompts.0stars
llm-oracleweiseer/llm-oracleUp-to-date LLM pricing + availability catalog with cited sources for AI agent routing.0stars