xluxx-mcp-server
v0.4.0
Published
XLUXX Trust Layer MCP Server — 16 tools for runtime trust scoring of 15,000+ MCP servers, Resonance Engine analysis, and Context Gate integrity checks. Installable in Claude Desktop, Cursor, and any MCP-compatible host.
Maintainers
Readme
XLUXX MCP Server
XLUXX Trust Layer as an MCP server. Gives any MCP-compatible host (Claude Desktop, Cursor, Continue, etc.) runtime access to trust scores, the Resonance Engine, and the Context Gate.
Install
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"xluxx": {
"command": "npx",
"args": ["-y", "xluxx-mcp-server"]
}
}
}With an API key (higher rate limits)
{
"mcpServers": {
"xluxx": {
"command": "npx",
"args": ["-y", "xluxx-mcp-server"],
"env": { "XLUXX_API_KEY": "your_key_here" }
}
}
}Get a free key at POST https://api.xluxx.net/register.
Direct CLI
npx -y xluxx-mcp-serverThe server speaks MCP over stdio.
Available tools (16)
Core trust tools
resolve_tool
Find the most reliable MCP tool for a given intent. Returns best server, confidence, resonance score, risk flags, and fallback.
intent(string, required) — what the agent needs to docategory(string, optional) — category filter
rank_tools
Rank a list of MCP tools by trust score with resonance metrics.
tools(string[], required) — tool/server IDs
recommend_toolchain
Get tool-chain recommendations for a complex multi-step task.
task(string, required) — natural-language description
check_server_status
Real-time health status of a specific MCP server, including outage detection.
server_id(string, required)
get_reliability_trends
Historical reliability data for a server — 24h, 7d, all-time stats.
server_id(string, required)
search_servers
Search 15,000+ MCP servers by capability, name, or category.
query(string, required)
get_alerts
Active alerts for degraded or down MCP servers.
- (no arguments)
Resonance Engine
get_resonance_score
Fractal Reliability Index and Coherence Drift for a server.
server_id(string, required)
analyze_toolchain_risk
Cumulative risk of a multi-step toolchain. Returns chain resonance, weakest links, and risk-adjusted reliability.
tools(string[], required) — ordered list of server/tool IDs
predict_reliability
Predict near-term reliability based on Fractal Reliability Index and recent trends.
server_id(string, required)horizon_hours(number, optional) — default 24
compare_alternatives
Compare a primary server against alternatives across trust, resonance, latency, and risk flags.
server_id(string, required) — primaryalternatives(string[], optional) — explicit alternatives; otherwise auto-selected
detect_drift
Detect coherence drift — schema changes, response inconsistency, or behavioral shift.
server_id(string, required)window(string, optional) —24h,7d,30d(default7d)
generate_trust_report
Comprehensive trust report: scores, flags, drift, alternatives, recommendations.
server_id(string, required)format(json|markdown, optional)
find_fallback
Best fallback server for a primary tool, ranked by capability overlap and trust.
server_id(string, required)intent(string, optional) — bias the search
monitor_fleet
Aggregate health, alerts, drift, and risk across a set of servers.
servers(string[], required)
explain_score
Factor breakdown explaining a server's trust and resonance scores, with weights.
server_id(string, required)
Context Gate (context firewall)
verify_context_integrity
One-shot drift check — verify the agent's context hasn't drifted during a conversation. Call before returning critical responses.
current_context(string, required) — current stateoriginal_context(string, optional) — full original for deep comparisonoriginal_context_hash(string, optional) — SHA-256 of originalconversation_length(number, optional)
verify_context
Continuous session-based context-firewall check. Requires a session_id from POST /context-gate/register. Returns gate=PASS if intact, gate=BLOCK with the missing instructions if drift exceeds threshold.
session_id(string, required)current_context(string, required)turn_number(number, optional)
Example prompts
Once installed, ask Claude:
- "Which MCP tool is most reliable for searching the web right now?"
- "Check the status of the GitHub MCP server."
- "Recommend a toolchain for researching competitors."
- "Compare brave-search against fetch and tavily — which is best?"
- "Predict reliability for postgres-mcp over the next 24 hours."
- "Find a fallback for the slack MCP server."
- "Monitor my fleet: github, slack, postgres, sendgrid."
- "Detect coherence drift on the openai-mcp server over the last 7 days."
- "Has my context drifted since the start of this conversation?"
Environment variables
| Variable | Default | Description |
|----------|---------|-------------|
| XLUXX_API_URL | https://api.xluxx.net | Override the API base URL |
| XLUXX_API_KEY | (none) | API key for higher rate limits |
Links
- API: https://api.xluxx.net
- Website: https://xluxx.net
- GitHub: https://github.com/DrDMT-VR/xluxx-trust
- npm: https://www.npmjs.com/package/xluxx-mcp-server
License
Apache-2.0 — Pablo Solorzano Cohen
