mcpField
A
Abis Mcp Server

Abis Mcp Server

Remote
github.com/cijlabsDeveloper Tools1Updated 3w ago
Repository

About Abis Mcp Server

ABIS MCP Server gives Claude Code teams tools to monitor LLM behavioral drift and model stability. It detects unannounced version changes and offers behavioral scoring, threat assessment, and reporting.

Key features

  1. 01Monitor behavioral drift across 11 models
  2. 02View nine-score behavioral scorecards
  3. 03Detect unannounced model version changes
  4. 04Assess threats such as prompt injection
  5. 05Generate compliance and operations reports

Use cases

  1. 01Track model behavior over time
  2. 02Compare stability across models
  3. 03Investigate drift alerts and potential attacks
  4. 04Prepare an EU AI Act compliance report

MCP support

The MCP server exposes model drift, scorecard, threat assessment, and reporting tools to Claude Code. Tool availability varies by access tier, and the free tier requires no API key.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://abis-mcp-production.up.railway.app/sse
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

Tools

Details

From the project README

ABIS Logo

ABIS MCP Server

AI Behavioral Drift Monitoring for Claude Code

GitHub Stars npm version License Node.js Documentation

27 tools · 11 models · Free / Pro / Business · Real-time drift detection · Patent-backed


What is ABIS?

ABIS (Adaptive Behavioral Intelligence System) monitors LLM behavioral drift in real-time. It detects when AI models silently change — shifts in tone, reasoning depth, safety calibration, or capability — even when providers don't announce updates.

This MCP server gives Claude Code native access to:

  • Live drift scores across 11 major models (GPT-4o, Claude, Gemini, DeepSeek)
  • 9-score behavioral scorecards (drift, coherence, entropy, alignment, and more)
  • Silent version change detection — catches updates providers don't announce
  • Immune threat assessment — prompt injection, coordinated attacks, organic drift
  • Correction simulation — test fixes before applying them
  • Compliance reporting — EU AI Act, SOC 2, and investor briefs

Backed by Patent GB2521226.7 · Built by CIJ Labs

Quick Start

One-line install (Claude Code)
/plugin install abis-drift-monitor
Manual setup

Add to your .mcp.json:

{
  "mcpServers": {
    "abis": {
      "type": "sse",
      "url": "https://abis-mcp-production.up.railway.app/sse"
    }
  }
}

No API key required for the Free tier. Connects to the ABIS cloud instantly.

Access Tiers

TierToolsKey RequiredWhat you get
Free15NoneDrift scores, scorecards, version changes, reports
Pro27MCP_PRO_KEY+ Scoring pipeline, memory analysis, immune system, correction simulation
Admin37ABIS_ADMIN_KEY+ Benchmarking, stress tests, config changes, pipeline stats
Pro/Admin connection
{
  "mcpServers": {
    "abis": {
      "type": "sse",
      "url": "https://abis-mcp-production.up.railway.app/sse",
      "headers": {
        "Authorization": "Bearer YOUR_KEY_HERE"
      }
    }
  }
}

37 Tools

Free Tier (15 tools)
ToolDescription
abis_ambient_statusQuick snapshot of all models — call at session start
abis_get_drift_scoresLive drift scores for all 11 monitored models
abis_get_scorecardFull 9-score breakdown (drift, coherence, entropy, alignment, etc.)
abis_whisperContextual status line (clear / monitor / alert / critical)
abis_drift_gatePre-response drift interceptor (Layer 3)
abis_observePost-response observer — how ABIS learns (Layer 3)
abis_historical_driftDrift time-series by model or date
abis_version_changesAll detected silent version changes across models
abis_memory_versionsSilent version change detection history per model
abis_memory_compareCross-model stability ranking
abis_corrector_strategiesList all 8 correction strategies with severity levels
abis_nervous_statusEvent bus health, registered handlers, recent event log
abis_monitoring_configCurrent phase, tier assignments, cost estimate
abis_ears_statusEARS webhook sources: Datadog, PagerDuty, Zendesk, Intercom
abis_generate_report5 report types: daily_summary, drift_alert, migration_report, compliance_doc, investor_brief
Pro Tier (+12 tools)
ToolDescription
abis_score_messageScore any text through the behavioral pipeline
abis_conversation_historyScored conversation log with drift per turn
abis_memory_profileFull 30-day behavioral memory for a model
abis_memory_trendTime-series trend for any score dimension
abis_immune_assessClassify threat: CLEAN / ORGANIC_DRIFT / PROMPT_INJECTION / COORDINATED_ATTACK
abis_immune_scanHistorical attack scan for a model
abis_simulateTest a correction before applying it
abis_corrector_historyWhich strategies work best per model
abis_cas_trendsWhat features drive drift (energy distance, entropy, coherence)
abis_cost_reportEstimated monthly API cost breakdown by model and tier
abis_cost_vs_drift30-day cost vs drift matrix
abis_run_benchmarkSingle or full benchmark run
Operator tools (internal only — not available on any customer plan)
ToolDescription
abis_api_cost_analysisMost-called models with drift trajectory over 7 days
abis_ears_correlationsDrift vs business impact correlations
abis_sdk_sessionsView SDK customer session history
abis_notification_historyRecent alert delivery log
abis_notification_configGet/set alert preferences (email, Slack, webhook)
abis_pipeline_statsIngestion statistics, vector counts
abis_training_statusXGBoost AUC scores, retraining history
abis_baseline_recalibrateRecompute drift baselines for all or one model
abis_monitoring_phaseSwitch phase (BASELINE / AGGRESSIVE / SUSTAIN / SCALE)
abis_stress_benchmark30 adversarial prompts across all models

Included Skills

After installing, you get two built-in skills:

  • /abis-ops — Daily health check: drift snapshot, version changes, stability ranking, cost summary
  • /abis-alerts — Investigate a drift alert: scorecard deep-dive, CAS trends, threat classification, correction recommendation

Models Monitored

ProviderModels
OpenAIgpt-4o, gpt-5.2, gpt-5.2-instant
Anthropicclaude-opus-4.6, claude-sonnet-4.5, claude-haiku-4.5
DeepSeekdeepseek-v3.2, deepseek-r1
Googlegemini-2.5-pro, gemini-2.5-flash, gemini-3-flash

Try It

"What is the current drift score for GPT-4o?"
"Run ABIS daily check"
"Has Claude Sonnet drifted in the last 7 days?"
"Generate an EU AI Act compliance report"
"Which model is most stable right now?"
"Have any models had silent version updates this week?"
"Show me the cost vs drift matrix"

Self-Hosted Setup

git clone https://github.com/cijlabs/abis-mcp
cd abis-mcp
npm install
cp .env.example .env
# Set ABIS_ADMIN_KEY, MCP_PRO_KEY, and RAILWAY_URL in .env
node index.js

Then add to Claude Code:

{
  "mcpServers": {
    "abis": {
      "command": "node",
      "args": ["/path/to/abis-mcp/index.js"],
      "env": {
        "ABIS_ADMIN_KEY": "your_admin_key_here"
      }
    }
  }
}

Architecture

abis-mcp/
├── index.js          # MCP server — SSE + stdio dual transport
├── config.js         # Environment config
├── db.js             # MongoDB persistence
├── api.js            # Railway API proxy with circuit breaker
├── tools/            # 37 tool implementations (one per file)
└── skills/           # /abis-ops and /abis-alerts
  • Dual transport: SSE (HTTP for cloud) + stdio (Claude Desktop/Code)
  • Rate limiting: 60 req/min global; per-tier limits (free=30, pro=120, admin=300)
  • Backend: Proxies to ABIS FastAPI on Railway
  • Persistence: MongoDB Atlas (sessions, scores, conversations, benchmarks)

Contributing

Contributions welcome! Please open an issue first to discuss what you'd like to change.

npm run dev    # Start with --watch on port 3100

License

MIT — see LICENSE for details.

About

CIJ Labs Ltd · Company No. 16984759 · London Patent: GB2521226.7 · cijlabs.com

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.