pAIchart MCP Hub

Get AI-powered recommendations and team performance analytics.

WHEN TO USE: Yes: Get optimization suggestions, assess risks, analyze team velocity and trends No: Reading POV/task data (use project), making changes (use perform)

ACTIONS: recommendations.get - AI optimization, risk, and workflow suggestions team.performance - Team velocity, completion rates, trend analysis

EXAMPLES: analytics({ action: "recommendations.get", povId: "cm3xyz", type: "RISK_MITIGATION" }) analytics({ action: "team.performance", timeframe: "30d", povId: "cm3xyz" })

RECOMMENDATION TYPES: AUTOMATION, OPTIMIZATION, RISK_MITIGATION, WORKFLOW_IMPROVEMENT, RESOURCE_ALLOCATION IMPACT LEVELS: CRITICAL, HIGH, MEDIUM, LOW TIMEFRAMES: 7d, 30d, 90d, 1y

WORKFLOW:

1. project(action: "pov.list") -> Find POV

2. analytics(action: "recommendations.get", povId: "...") -> Get suggestions

3. analytics(action: "team.performance", povId: "...") -> Review team metrics

4. perform(action: "task.create", ...) -> Act on recommendations

RETURNS: recommendations.get: { recommendations: [{ type, title, impact, effort, actions }], total, _meta } team.performance: { metrics: { velocity, completionRate, trends }, _meta }

SEE ALSO: project - Query POVs and tasks perform - Execute task actions

Retrieve detailed information for a specific resource including full content and metadata. Use the ID from search results (format: type-id, e.g., 'pov-123', 'task-456').

WHEN TO USE: ✅ Have resource ID from search results ✅ Need quick condensed view (50KB limit) ✅ Cross-resource fetch (POVs, tasks, executions, templates) ❌ Need full POV details with team IDs (use project(action: "pov.details") instead) ❌ Need to search first (use search then fetch)

EXAMPLES: • fetch("pov-cmgalshus00bcyx39sfdutido") → POV summary • fetch("task-cmgalshy500fayx39xe8f4xn1") → Task with execution history • fetch("template-cmf6gvbkl0005yxvvqayf35ug") → Agent template details

WORKFLOW:

1. search("keywords") → Find resources

2. fetch("type-id") → Get condensed details

3. project(action: "pov.details", povId: "...") → Get full details if needed

RETURNS: Format: {id, title, text, url, metadata} Note: fetch returns direct object, search returns wrapped {results: [...]}

SEE ALSO: • project - Comprehensive POV data with team member IDs • search - Find resources first

Search built-in prompt templates for common Hub workflows like service registration, discovery, and orchestration.

WHEN TO USE: ✅ Learning Hub workflows - see example prompts ✅ Finding orchestration templates ✅ Discovering service lifecycle prompts ✅ Getting started with multi-service workflows ❌ Listing MCP services (use services(action: "discover"))

EXAMPLES: • list_prompts() → All available prompts • list_prompts(query: 'orchestrate') → Workflow orchestration prompts • list_prompts(domain: 'POV') → POV-specific prompts • list_prompts(category: 'WORKFLOW') → Workflow templates only

SEARCH PARAMETERS: • query - Free-text search across name and content • domain - Filter by domain (POV, tasks, agents, general) • category - Filter by type (WORKFLOW, ANALYSIS, ONBOARDING)

RETURNS: • prompts - Array of prompt templates with name, description, category • total - Count of matching prompts • suggestions (if query used) - Related prompts you might want

PROMPT DOMAINS: • POV - Proof of Value management prompts • tasks - Task management workflows • agents - Agent execution prompts • general - Hub and service management

SEE ALSO: • prompt_command - Execute a prompt template • registry(action: "list") - Check your identity and services • services(action: "discover") - Find available services

Execute task management operations across 13 actions in 6 categories.

WHEN TO USE: Yes: Creating/updating tasks, assigning agents, executing agents, marking tasks complete No: Reading data (use project), AI insights (use analytics), external services (use services)

[WHICH ACTION DO I USE?]

Want to CREATE something? pov.create - New POV with team and phases (ADMIN ONLY!) task.create - New task (povId REQUIRED!) stage.create - New stage in phase

Want to MODIFY a task? task.update - Change ANY field (status, assignee, priority, title, description, dueDate) task.assign - Change assignee only task.complete - Mark done task.comment - Add comment

Want to use AGENTS for automation?

1. agent.assign - Attach template to task (required first step!)

2. agent.configure - Optional customization

3. agent.execute - Run the agent

4. agent.status - Check if still running

5. agent.results - Get output and artifacts

Want ANALYTICS? analytics.generate - Generate performance reports

EXAMPLES: perform({ action: "task.create", parameters: { povId: "cm3xyz", title: "New Task" } }) perform({ action: "task.update", taskId: "cm3abc", status: "COMPLETED" }) perform({ action: "agent.execute", taskId: "cm3abc" })

WORKFLOW:

1. project(action: "pov.list") -> Find POV

2. project(action: "task.list", povId: "...") -> Find task

3. perform(action: "task.update", ...) -> Make changes Agent workflow: template(action: "list") -> perform(agent.assign) -> perform(agent.execute) -> perform(agent.results)

RETURNS: task.*: { success, task: { id, title, status }, _meta } agent.execute: { success, executionId, status: "RUNNING", _meta } agent.results: { success, artifacts: [...], output, _meta }

SEE ALSO: project - Query POVs and tasks before making changes analytics - AI recommendations and team metrics services - Call external services from workflows template - Agent template management (ADMIN)

Query POVs (Proof of Value projects), tasks, and task context.

WHEN TO USE: Yes: Browse POVs by status/geography, get team member IDs, list tasks, deep-dive task context No: Making changes (use perform), searching by keyword (use search), getting AI insights (use analytics)

ACTIONS: pov.list - List POVs with filtering (status, geography, customer) pov.details - Get comprehensive POV details (team IDs, phases, stages) task.list - List tasks with flexible filtering task.context - Deep dive on a specific task (execution history, dependencies)

EXAMPLES: project({ action: "pov.list", status: "IN_PROGRESS", limit: 20 }) project({ action: "pov.details", pov_name: "BlackEye" }) project({ action: "task.list", pov_name: "BlackEye", status: "OPEN" }) project({ action: "task.context", taskId: "cm123...", includeHistory: true })

WORKFLOW:

1. project(action: "pov.list") -> Browse POVs

2. project(action: "pov.details", povId: "...") -> Get team/phase IDs

3. project(action: "task.list", povId: "...") -> See tasks

4. project(action: "task.context", taskId: "...") -> Deep dive

5. perform(action: "task.update", ...) -> Make changes

RETURNS: pov.list: { povs: [...], total, _meta } pov.details: { pov: { id, title, status, team: [{id, name, role}], phases: [...] }, _meta } task.list: { tasks: [...], total, _meta } task.context: { task: { id, title, status, history: [...], dependencies: [...] }, _meta }

SEE ALSO: perform - Create, update, execute operations analytics - AI recommendations and team metrics search - Natural language resource discovery template - Agent template management (ADMIN)

Execute MCP prompt templates for guided workflows and automation.

WHEN TO USE: ✅ Execute workflow guidance prompts (orchestration, service registration) ✅ Run templated operations with pre-built prompts ✅ Discover available prompts with /prompt list ✅ Get help with /prompt help ❌ General search across resources (use search instead) ❌ Listing specific services (use services(action: "discover"))

EXAMPLES: • prompt_command(command: '/prompt list') → List all available prompts • prompt_command(command: '/prompt help') → Show usage guide • prompt_command(command: '/prompt select_pov search_term="BlackEye"') → Find POV by name

COMMON PROMPTS: • select_pov - Find and select a POV by name • show_available_prompts - List all prompts with descriptions

COMMAND FORMAT: • Basic: /prompt [name] • With args: /prompt [name] key=value key2="value with spaces" • List: /prompt list • Help: /prompt help

WORKFLOW:

1. list_prompts() → Discover available prompts

2. prompt_command(command: '/prompt [name]') → Execute prompt

3. Follow prompt output for next steps

4. Use perform(action: "task.create", ...) for any resulting tasks

SEE ALSO: • list_prompts - Search prompts by domain/category • perform - Create tasks from prompt guidance • services(action: "discover") - Find available services

Register, update, delete, and inspect your MCP services in the hub.

WHEN TO USE: Yes: Register a new service, list/update/delete your services, inspect service tool schemas No: Discovering other services (use services action: "discover"), calling services (use services action: "call")

ACTIONS: register - Register a new MCP service list - List your registered services and identity update - Update service configuration delete - Permanently delete a service (GDPR Right to Erasure) tools - Get detailed tool definitions for a service

EXAMPLES: registry(action: 'register', name: 'my-api', endpoint: 'https://api.example.com', category: 'data-services', capabilities: {tools: ['fetch', 'query']}) registry(action: 'list') registry(action: 'list', status: 'ACTIVE', includeMetrics: true) registry(action: 'update', service_name: 'my-api', updates: {endpoint: 'https://new.api.com'}) registry(action: 'delete', service_name: 'my-api', confirm: true) registry(action: 'tools', service_name: 'notification-service')

WORKFLOW:

1. registry(action: 'register', ...) - Register your service

2. services(action: 'discover') - Others find your service

3. registry(action: 'tools', service_name: '...') - Others discover your tool parameters

4. services(action: 'call', ...) - Others use your tools

RETURNS: register: { service: { id, name, status }, _meta } list: { user: { email, role }, services: [...], total, _meta } update: { service: { id, name, updated fields }, _meta } delete: { deleted: true, serviceName, _meta } tools: { service: { name, version }, tools: [{ name, description, inputSchema }], toolCount, _meta }

SEE ALSO: services - Discover, call, and orchestrate services search - Find resources by keyword

Search across Projects (Proof of Value), tasks, agent activities, and templates for comprehensive research. Returns results in ChatGPT-compatible format for natural language queries.

WHEN TO USE: ✅ Natural language discovery across all resource types ✅ Don't know exact POV/task name but know keywords ✅ Quick exploration ("find email gateway projects") ❌ Need structured filters (use project instead) ❌ Already have resource ID (use fetch instead)

EXAMPLES: • search("CyberDefense") → All CyberDefense resources (POVs, tasks, templates) • search("validation security") → Security-related validation work • search("QA testing") → Testing tasks and templates

WORKFLOW:

1. search("keywords") → Discover resources

2. fetch("pov-xyz") → Get full details for interesting result

3. perform(action: "task.update", ...) → Make changes

RETURNS: Format: {results: [{id, title, url, ...}, ...]} Note: search returns wrapped object {results: [...]}, fetch returns direct object

SEE ALSO: • project - Structured POV/task filtering by status/region • fetch - Get details for specific resource ID

Discover, call, monitor, and orchestrate external MCP services through the Hub.

WHEN TO USE: Yes: Find services by capability, call service tools, check health, run multi-service workflows No: Registering/managing your own services (use registry), querying pAIchart data (use project)

ACTIONS: discover - Search Hub registry for services by capability/category call - Execute a tool on a registered service health - Check health status and metrics of a specific service workflow.execute - Run multi-service workflows (sequential, parallel, conditional) workflow.status - Check status of a workflow execution workflow.cancel - Cancel a running workflow workflow.list - List workflow execution history

EXAMPLES: services({ action: "discover", category: "monitoring" }) services({ action: "call", targetService: "sentry-mcp", tool: "create_issue", arguments: { title: "Bug" } }) services({ action: "health", service_name: "sentry-mcp" }) services({ action: "health", service_name: "sentry-mcp", realtime: true }) services({ action: "workflow.execute", workflowName: "daily-energy-weather" }) services({ action: "workflow.execute", steps: [...], executionMode: "sequential" }) services({ action: "workflow.status", executionId: "clxyz123abc" }) services({ action: "workflow.cancel", executionId: "clxyz123abc", reason: "No longer needed" }) services({ action: "workflow.list", status: "FAILED", limit: 10 })

CATEGORIES: ai-intelligence, data-services, automation, monitoring, communication, security

WORKFLOW:

1. discover → Find services by capability

2. health → Verify service is healthy before calling

3. call → Execute a tool on the service

4. workflow.execute → Chain multiple service calls

5. workflow.status → Monitor workflow progress

6. workflow.list → Review execution history

7. workflow.cancel → Stop a running workflow

RETURNS: discover: { services: [{ id, name, description, capabilities, status }], total, _meta } call: { result: , service, tool, _meta } health: { status, uptime, latency, metrics, _meta } workflow.execute: { executionId, status: "RUNNING", _meta } workflow.status: { executionId, status, steps: [{ status, result }], _meta }

SEE ALSO: registry - Register, update, delete your own services project - Query POV/task data directly (faster than service call for pAIchart data) perform - Action pattern reference (same action-based dispatching)

List and inspect agent templates. ADMIN access required.

WHEN TO USE: Yes: Browse available agent templates, inspect template configuration before assigning No: Assigning a template to a task (use perform action: "agent.assign")

ACTIONS: list - List all available agent templates details - Get full details for a specific template

EXAMPLES: template({ action: "list", limit: 50 }) template({ action: "details", templateId: "cm123..." }) template({ action: "details", template_name: "Security Analyst" })

WORKFLOW:

1. template(action: "list") -> Browse templates

2. template(action: "details", templateId: "...") -> Inspect config

3. perform(action: "agent.assign", taskId: "...", agentTemplateId: "...") -> Assign to task

4. perform(action: "agent.execute", taskId: "...") -> Run agent

RETURNS: list: { templates: [{ id, name, description, role }], total, _meta } details: { template: { id, name, systemPrompt, tools, config }, _meta }

SEE ALSO: perform(action: "agent.assign") - Attach template to a task perform(action: "agent.execute") - Run an agent after assignment