Atlas / MCP servers / tosin2013 / ADR Analysis

ADR AnalysisBLOCK

Name: ADR Analysis
Author: tosin2013

mcp/tosin2013/adr-analysis

MCP server by tosin2013 · adr · ai agents · architectural decision records · Audited 2026-10-08

Your ADRs are lying to you. MCP server with live drift detection, content safety, and decision memory — validates architectural decisions against your actual code.

Verdict

BLOCK

Grade

Trust score

37 /100

Exposed tools

115 101r · 13w · 1d

Transport

sse · stdio

License

MIT

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

[](https://github.com/tosin2013/mcp-adr-analysis-server) [](LICENSE) [](https://www.npmjs.com/package/mcp-adr-analysis-server) [](https://nodejs.org/) [](https://www.typescriptlang.org/) [](https://github.com/tosin2013/mcp-adr-analysis-server/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22)

Your ADRs are lying to you. This MCP server catches it — live drift detection validates architectural decisions against your actual code. Plus content safety, decision memory, and 63 tools powered by your host LLM via CE-MCP.

What is MCP?
Prerequisites
Quick Installation
Quick Setup
Usage Examples
Use Cases
Technology Stack
Project Structure
Testing
ADR Aggregator Integration
Development
Troubleshooting
Security & Performance
Contributing
Resources
License

What is MCP?

The Model Context Protocol (MCP) is an open standard that enables seamless integration between AI assistants and external tools and data sources. Think of it as a universal adapter that lets AI assistant

Read from source at commit 1060b7a805c7OBSERVED · 2026-10-08

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code (npm)

claude mcp add mcp-adr-analysis-server --env OPENROUTER_API_KEY=${OPENROUTER_API_KEY} -- npx -y [email protected]

Exposed tools (115)

101 read · 13 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`CQRS`	read	Command Query Responsibility Segregation
`MVC`	read	Model-View-Controller
`Microservices`	read	Distributed architecture
`MongoDB`	read	Document database with horizontal scaling capabilities
`Node.js`	read	Backend runtime
`Observer`	read	Observer pattern
`PostgreSQL`	read	Relational database system
`React`	read	UI Library
`Redis`	read	In-memory data store
`Test`	read	Test
`action`	read	Description of the action to be performed
`action_confirmation`	read	Confirm actions before writing files to disk
`adrDirectory`	read	Directory containing ADR files
`ambiguity_resolution`	read	Resolve ambiguities in project analysis or requirements
`ambiguous_items`	read	List of ambiguous items that need clarification
`analysisDepth`	read	Analysis depth: shallow, deep
`analyze_adr_timeline`	read	Analyze ADR timeline with smart time tracking, adaptive thresholds, and actionable recommendations. Auto-detects project context (startup/growth/mature) and generates prioritized work queue based on staleness, implementation lag, and technical debt.
`analyze_content_security`	read	Analyze content for sensitive information using AI-powered detection with optional memory integration for security pattern learning
`analyze_deployment_progress`	read	Analyze deployment progress and verify completion with outcome rules
`analyze_environment`	read	Analyze environment context and provide optimization recommendations with optional memory integration for environment snapshot tracking
`analyze_gaps`	read	Scan local codebase and compare with ADRs to detect bi-directional gaps. Finds: (1) ADR-to-code gaps: file references in ADRs that do not exist, (2) Code-to-ADR gaps: technologies in package.json and architectural patterns without ADR coverage. Reports gaps to ADR Aggregator for tracking.
`analyze_project_ecosystem`	read	Comprehensive recursive project ecosystem analysis with advanced prompting techniques (Knowledge Generation + Reflexion)
`apply_basic_content_masking`	write	Apply basic content masking (fallback when AI is not available)
`architecture`	read	Architecture rules
`args`	read	Analysis arguments including code patterns and context
`baseline_analysis`	read	Generate comprehensive baseline analysis for existing projects
`bootstrap_validation_loop`	read	Bootstrap self-learning validation and deployment workflow
`compare_adr_progress`	read	Compare TODO.md progress against ADRs and current environment to validate implementation status
`configure_custom_patterns`	read	Configure custom sensitive patterns for a project
`configure_output_masking`	read	Configure content masking for all MCP outputs
`content_type`	read	Type of content being created (code, documentation, configuration)
`context`	read	Context where ambiguity was detected
`create_research_template`	write	Create a research template file for documenting findings
`create_rule_set`	write	Create machine-readable rule set in JSON/YAML format
`custom_rule_definition`	read	Define custom architectural rules and validation criteria
`deployment_readiness`	write	Comprehensive deployment readiness validation with test failure tracking, deployment history analysis, and hard blocking for unsafe deployments. Integrates with smart_git_push for deployment gating.
`discover_existing_adrs`	read	Discover and catalog existing ADRs in the project
`estimationType`	read	Type of estimation: effort, duration, resources
`existing_rules`	read	Existing rules in the project (JSON array)
`expand_analysis_section`	read	Retrieve full analysis content from tiered responses. Expand entire analysis or specific sections stored in memory. Use this when a tool returns a summary with an expandable ID.
`expand_memory`	read	Phase 3: Retrieve and expand stored content from a tiered response using its expandable ID
`files`	read	List of files that will be affected (JSON array)
`filters`	read	Filters for list action (JSON array)
`focus_areas`	read	Specific areas to focus on (comma-separated)
`generate_adr_bootstrap`	read	Bootstrap ADR structure for a project
`generate_adr_from_decision`	read	Generate a complete ADR from decision data. TIP: Reference @.mcp-server-context.md to align with existing architectural patterns and decisions.
`generate_adrs_from_prd`	read	Generate Architectural Decision Records from a Product Requirements Document with advanced prompting techniques (APE + Knowledge Generation)
`generate_content_masking`	read	Generate masking instructions for detected sensitive content
`generate_deployment_guidance`	read	Generate deployment guidance and instructions from ADRs with environment-specific configurations
`generate_research_questions`	write	Generate context-aware research questions and create research tracking system
`generate_rules`	read	Generate architectural rules from ADRs and code patterns
`get_adr_context`	read	Fetch ADR context from ADR Aggregator including summaries, diagrams, timeline data, and code links. Useful for getting a consolidated view of architectural decisions.
`get_adr_diagrams`	read	Get Mermaid diagrams for ADRs from ADR Aggregator. Includes workflow, relationship, and impact diagrams. Requires Pro+ tier.
`get_adr_priorities`	read	Get ADR priorities for roadmap and backlog planning from ADR Aggregator. Returns prioritized ADRs with scores, dependencies, blockers, implementation status, and gap counts.
`get_adr_templates`	read	Get domain-specific ADR templates and anti-patterns from ADR Aggregator. Includes best practices for web applications, microservices, APIs, and more. No authentication required.
`get_architectural_context`	read	Get detailed architectural context for specific files or the entire project, automatically sets up ADR infrastructure if missing, and provides outcome-focused workflow for project success
`get_conversation_snapshot`	read	Phase 3: Get current conversation context snapshot for resumption or analysis
`get_development_guidance`	read	Get comprehensive development guidance that translates architectural decisions and workflow recommendations into specific coding tasks, implementation patterns, and development roadmap
`get_gaps`	read	Get current code gaps from ADR Aggregator. Returns gaps with their status (open, dismissed, resolved) for tracking and management.
`get_knowledge_graph`	read	Get cross-repository knowledge graph from ADR Aggregator with analytics and insights. Visualize ADR relationships across repositories. Requires Team tier.
`get_memory_stats`	read	Phase 3: Get statistics about stored conversation memory
`get_server_context`	read	Generate a comprehensive context file showing the server
`get_staleness_report`	read	Get ADR staleness report from ADR Aggregator with review compliance metrics. Identifies stale ADRs that need attention and provides governance insights.
`get_workflow_guidance`	read	Get intelligent workflow guidance and tool recommendations based on your goals and project context to achieve expected outcomes efficiently
`goal_specification`	read	Specify project goals and requirements for comprehensive analysis
`granularity`	read	Estimation granularity: high, medium, low
`includeDownstream`	read	Include downstream dependencies (true/false)
`includeUncertainty`	read	Include uncertainty analysis (true/false)
`includeUpstream`	read	Include upstream dependencies (true/false)
`incorporate_research`	read	Incorporate research findings into architectural decisions
`interactive_adr_planning`	read	Interactive guided ADR planning and creation tool - walks users through structured decision-making process with research integration, option evaluation, and automatic ADR generation. TIP: Start by reading @.mcp-server-context.md to understand project context and previous decisions.
`manage_cache`	destructive	Manage MCP resource cache (clear, stats, cleanup)
`mcp_planning`	read	Enhanced project planning and workflow management tool - phase-based project management, team resource allocation, progress tracking, risk analysis, and executive reporting
`memory_loading`	read	Advanced memory loading tool for the memory-centric architecture. Query, explore, and manage memory entities and relationships. Load ADRs into memory system and perform intelligent queries.
`newStatus`	write	New status for update action
`perform_research`	read	Perform research using cascading sources: project files → session/tool-usage tracker → environment resources → web search (fallback)
`phase`	read	Development phase: both, test, production
`preview`	read	Preview of changes to be made
`projectContext`	read	Project context for technology detection
`projectPath`	read	Path to the project directory
`project_path`	read	Path to the project directory
`project_type`	read	Type of project (web, mobile, api, library, tool, etc.)
`query_conversation_history`	read	Phase 3: Search and retrieve conversation sessions based on filters
`release_tracking`	read	Track releases mapped to ADR decisions. Generates changelogs, manages milestones, compares releases, and assesses release readiness. Supports greenfield and brownfield projects. Writes CHANGELOG.md, creates GitHub Releases and Milestones.
`request_action_confirmation`	read	Request confirmation before applying research-based changes
`review_existing_adrs`	read	Review existing ADRs against actual code implementation with cloud/DevOps expertise. TIP: After review, call get_server_context to update @.mcp-server-context.md with findings.
`rule_category`	read	Category of rule (architectural, coding, security, performance, documentation)
`scale_requirements`	read	Expected scale and performance requirements
`scope`	read	Task scope: all, pending, in_progress
`search_codebase`	read	Atomic tool for searching codebase files based on query patterns. Returns raw file matches with relevance scores. Extracted from ResearchOrchestrator per ADR-018.
`search_tools`	read	[DEPRECATED host-native, ADR-023] Search and discover available tools by category, keyword, or capability. Use this to find the right tool for a task without loading all tool schemas. Returns lightweight tool metadata by default; use includeSchema:true for full schemas.
`secret_prevention_guidance`	read	Proactive guidance to prevent secret exposure in code and documentation
`set_project_path`	write	Dynamically set the active project path for the current session. Call this at the start of a session to switch between projects without restarting the server or modifying environment variables. All subsequent tool calls will use this path as the default.
`smart_git_push`	write	AI-driven security-focused git push with credential detection, file filtering, and deployment metrics tracking. Tests should be run by calling AI and results provided.
`smart_score`	write	Central coordination for project health scoring system - recalculate, sync, diagnose, optimize, and reset scores across all MCP tools
`suggest_adrs`	read	Suggest architectural decisions with advanced prompting techniques (Knowledge Generation + Reflexion). TIP: Read @.mcp-server-context.md first for project history, patterns, and previous ADRs to ensure consistency.
`suggestions`	read	Suggested interpretations or options
`sync_to_aggregator`	write	Sync ADRs to ADR Aggregator platform (https://adraggregator.com) for centralized tracking, visualization, and team collaboration. Supports incremental and full sync modes with optional metadata.
`target_audience`	read	Target audience or users of the project
`target_location`	read	Where the content will be stored/committed
`taskId`	write	Specific task ID (for update/complete actions)
`test`	read	Test profile
`todo_dependency_analysis`	read	Analyze task dependencies and critical path optimization
`todo_estimation`	read	Provide accurate task estimation and timeline planning
`todo_status_management`	read	Manage task status, priorities, and progress tracking
`todo_task_generation`	read	Generate comprehensive development task list from ADRs with cloud/DevOps expertise
`tool_chain_orchestrator`	read	AI-powered dynamic tool sequencing - intelligently analyze user requests and generate structured tool execution plans
`troubleshoot_guided_workflow`	read	Structured failure analysis and test plan generation with memory integration for troubleshooting session tracking and intelligent ADR/research suggestion capabilities - provide JSON failure info to get specific test commands
`update_implementation_status`	write	Update the implementation status of synced ADRs directly from the IDE. Supports statuses: not_started, in_progress, implemented, deprecated, blocked. Requires Pro+ tier.
`update_knowledge`	write	ADR-018: Simple CRUD operations for project session state. Not a graph database — keyword retrieval over local JSON snapshots. Add/remove entities (intents, ADRs, tools, code) and relationships. Use knowledge://graph resource to read current state (zero token cost).
`validate_adr`	read	Validate an existing ADR against actual infrastructure reality using research-driven analysis. TIP: Compare findings against patterns in @.mcp-server-context.md for consistency checks.
`validate_adr_compliance`	read	Validate ADR compliance against implementation via ADR Aggregator. Checks that code actually implements documented decisions. Requires Pro+ tier.
`validate_all_adrs`	read	Validate all ADRs in a directory against actual infrastructure reality
`validate_content_masking`	read	Validate that content masking was applied correctly
`validate_rules`	read	Validate code against architectural rules

Trust audit

BLOCKgrade F · trust 37/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

Layer	What it checks	Result
L0	Provenance & inventory	WARN
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	FAIL
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (6 observation(s))
Network: declared (6 observation(s))
Shell: declared (5 observation(s))
Dependencies: not all pinned
Secrets in source: found

Findings (25)

HIGHFilesystem / path · fs.credential_store · CWE-22, CWE-59

src/utils/gitleaks-detector.ts:375

/id_rsa$/,

Why it matters. touches a credential store

HIGHFilesystem / path · fs.credential_store · CWE-22, CWE-59

src/utils/gitleaks-detector.ts:376

/id_ed25519$/,

Why it matters. touches a credential store

HIGHPrompt injection · prompt.tool_poisoning · CWE-94, CWE-1427

docs/how-to-guides/prompting-guide.md:413

"As an enterprise architect, I need to assess [SYSTEM] for compliance and modernization. Please use analyze_project_ecosystem with enterprise focus, then suggest_adrs for governance patterns, and gene

Why it matters. a tool description carrying instructions to the agent

Fix. tool descriptions describe the tool; nothing else

MEDIUMInventory / provenance · inv.binary · CWE-1104

.npm-cache/_cacache/content-v2/sha512/03/aa/7fa6effa7f205c0354d1cff1aa5983d58a996b7ed716da0642f6aefd9e0342280791fd7de070475740797828d5d5fb7c20209d423e4250dc81ccea572cc8

7fa6effa7f205c0354d1cff1aa5983d58a996b7ed716da0642f6aefd9e0342280791fd7de070475740797828d5d5fb7c20209d423e4250dc81ccea572cc8