MCP For Claude Marketplace
Silent Change Detector

Silent Change Detector MCP Connector for Claude

A+

Identifies unauthorized or unrequested code modifications by comparing file state snapshots.

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

Silent Change Detector monitors AI agent execution cycles to identify unauthorized or unrequested code modifications. By capturing a cryptographic SHA-256 state snapshot before an agent turn and comparing it to the state after, it classifies changes as intended, cascading, or unintended. This tool helps users detect 'silent' changes where agents refactor code or align with outdated documentation without explicit permission. Use snapshot_state to establish a baseline, detect_changes to classify modifications, and analyze_documentation_drift to find discrepancies between code and documentation.

gitsha256ai-agentscode-integrityaudit

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

analyze_documentation_drift

Detects if code was modified to align with incorrect or outdated documentation

detect_changes

Compares a previous snapshot against the current state to identify changes and their relation to user intent

snapshot_state

Captures the current state of the workspace to establish a baseline for comparison

See how to talk to your AI agent using Silent Change Detector.

Take a snapshot of my current workspace.

Snapshot created successfully with ID: snap_12345.

Check if any unintended changes were made after my last instruction to fix the typo in main.ts.

No unintended changes detected. The only modification was in main.ts, which matches your request.

Did the agent change any code to match the outdated documentation in docs/api.md?

Drift detected in src/utils.ts. The agent modified the logic to match the outdated documentation provided.

A silent change is any modification categorized as unintended or any change that occurs without a corresponding instruction in the user's prompt.

Related Connectors

NEW
Estimate Scope Reconciliation

Estimate Scope Reconciliation MCP

4 tools Official

Detect discrepancies between project baselines and revised estimates to generate approval checklists.

A+ View details →
NEW
Policy Change Impact Brief

Policy Change Impact Brief MCP

4 tools Official

Maps textual policy revisions to insured assets and planned activities to prevent coverage gaps.

A+ View details →
Agent Self-Reflection Sentiment Scanner

Agent Self-Reflection Sentiment Scanner MCP

3 tools Official

Analyzes agent execution logs to detect self-correction markers and measure stability.

A+ View details →
Observability & Tracing Calculator

Observability & Tracing Calculator MCP

3 tools Official

Deterministic engine for distributed tracing metrics, cost estimation, and anomaly detection.

A+ View details →