MCP For Claude Marketplace
Agent Config Drift Detector

Agent Config Drift Detector MCP Connector for Claude

A+

Detects unauthorized changes to agent configurations by comparing SHA-256 hashes.

3 tools Official Updated Oct 1, 2026 Official Vinkius Partner
More Details Connect to Claude

This MCP server ensures the integrity of LLM agent configurations by detecting 'drift'--unintended or unauthorized modifications to critical parameters. It uses deterministic SHA-256 hashing to compare current configurations against a known good baseline. By monitoring fields like system prompts and permissions, it identifies specific deviations and calculates a risk severity score. Use check_config_integrity to detect changes, get_drift_report for granular diffs, and evaluate_risk_level to assess the impact of detected drift.

securityintegritydrift-detectionsha256llm-ops

3 tools expose this connector's capabilities to your AI agent.

evaluate_risk_level

Calculates the severity of a detected drift based on the nature of the changes

get_drift_report

Provides a detailed breakdown of the specific differences between two versions of a configuration

check_config_integrity

Determines if the current agent configuration has deviated from its authorized baseline

See how to talk to your AI agent using Agent Config Drift Detector.

Check if the current configuration for agent 'auth-service-01' has drifted.

{"hasDrifted": true, "currentHash": "a1b2c3d4...", "driftSeverity": "HIGH", "changedFields": [{"field": "system_prompt", "oldVal": "...", "newVal": "..."}]}

What is the risk level for these changed fields: ['permissions', 'allowed_tools']?

{"severity": "CRITICAL", "riskScore": 95}

Get a detailed drift report for agent 'data-processor' between version v1.0 and v1.1.

{"fieldName": "allowed_tools", "type": "addition", "oldValue": "["read_db"]", "newValue": "["read_db", "write_db"]", "severityImpact": "LOW"}

The server computes a deterministic SHA-256 hash of critical configuration sections and compares it against a stored baseline. If the hashes differ, it performs a deep inspection to find the exact changed fields.

Related Connectors

Claude Conversation Drift Detector

Claude Conversation Drift Detector MCP

3 tools Official

Monitors AI agent focus by detecting task drift and topic shifts.

A+ View details →
test-regression-runner

test-regression-runner MCP

3 tools Official

Automated verification system that executes targeted test suites and detects regressions or flakiness.

A+ View details →
Agent Hallucination Cross-Checker

Agent Hallucination Cross-Checker MCP

3 tools Official

A deterministic engine to audit agent outputs by measuring consensus, source validity, and semantic contradictions.

A+ View details →
License Header Injector and Verifier

License Header Injector and Verifier MCP

3 tools Official

Enforce, audit, and repair mandatory copyright headers in source files.

A+ View details →