personizeai
9 verified skills
personize-solution-architect
How to design, build, evaluate, and evolve Personize integrations — including the advanced multi-step `instructions[]` patterns those integrations run. Covers situation assessment, entity modeling, property design for extraction quality, integration topology, workspace design, governance architecture, scaling, schema evolution, smart update for AI-powered guideline and schema evolution, evaluation and optimization of extraction and recall quality, infrastructure management (entity types, MCPs, destinations), project lifecycle, deployment patterns, and authoring advanced multi-step prompt patterns (conditional branching by tier, multi-source reconciliation, compliance-gated generation, persona fanout, bounded research with source triangulation, few-shot calibrated classification, checklist-gated workflows, self-reflective refinement loops). Use when designing solutions, creating schemas, planning integrations, evaluating extraction quality, updating or evolving guidelines, improving system performance, or debugging instruction chains that silently fail, produce inconsistent confidence outputs, loop indefinitely, or write bad data to records.
tools
personize-enabler
Ready-to-execute resources for building with Personize: TypeScript scripts (CSV import, CRM sync, outreach, monitoring), n8n workflow templates, collection schema presets, and governance guideline templates. Grab, customize, and run. Use whenever you need a script, template, starter schema, or pre-built workflow.
development
personize-agent-core
How to think, compose, coordinate, self-correct, and learn as an autonomous Personize-powered agent. Covers bootstrap, the core loop (recall-govern-act-store), tool composition, multi-agent coordination, self-correction, analysis patterns, code generation, governance compliance, risk management, resilience, learning, observability, and cost awareness. Use as the foundation for any Personize-powered AI agent on any platform.
tools
personize-reference
Complete lookup layer for every Personize API endpoint, SDK method, CLI command, and MCP tool — including schedules (recurring/one-time `run_prompt` and `send_notification`) and CRM passthrough (HubSpot/Salesforce direct API access via the org's managed OAuth connection). Exhaustive cross-interface operation tables, error handling, authentication, response schemas, and FAQ files for instant answers. Use whenever looking up how to call a specific operation, what parameters it takes, what errors to expect, how interfaces map to each other, how to schedule a recurring prompt, or how to call HubSpot/Salesforce REST APIs without managing OAuth credentials.
tools
personize-reference
Complete lookup layer for every Personize API endpoint, SDK method, CLI command, and MCP tool. Exhaustive cross-interface operation tables, error handling, authentication, response schemas, and FAQ files for instant answers. Use whenever looking up how to call a specific operation, what parameters it takes, what errors to expect, or how interfaces map to each other.
tools
personize-agent-workspace
Turn any record into a shared workspace where agents and humans collaborate. Attach a simple workspace schema to any entity — contacts, companies, deals, projects, tickets — and let any participant contribute updates, tasks, notes, and issues. Use this skill whenever the user wants multi-agent collaboration, shared context on an entity, agent handoffs, workspace-based coordination, or the three-layer agent operating model (Guidelines + Memory + Workspace). Also trigger when they mention multiple agents working on the same record, deal rooms, account intelligence, customer health monitoring, cross-functional coordination, or progressive autonomy for AI agents.
testing
personize-governance
Manages organizational guidelines, policies, and best practices as governance variables accessible to all AI agents via SmartContext. Use this skill whenever the user wants to create, update, or manage guidelines, brand voice, compliance policies, playbooks, ICPs, sales playbooks, tone rules, or any organizational rules. Also trigger when the user mentions smartGuidelines, governance variables, GitOps sync of policies, team knowledge sharing, AI agent rules, or when they want all their AI tools to follow the same policies. Even if they just say 'set up rules' or 'add a policy', this is the right skill.
tools
personize-architect
How to design, build, evaluate, and evolve Personize integrations — including the advanced multi-step `instructions[]` patterns those integrations run. Covers situation assessment, entity modeling, property design for extraction quality, integration topology, workspace design, governance architecture, scaling, schema evolution, smart update for AI-powered guideline and schema evolution, evaluation and optimization of extraction and recall quality, infrastructure management (entity types, MCPs, destinations), project lifecycle, deployment patterns, and authoring advanced multi-step prompt patterns (conditional branching by tier, multi-source reconciliation, compliance-gated generation, persona fanout, bounded research with source triangulation, few-shot calibrated classification, checklist-gated workflows, self-reflective refinement loops). Use when designing solutions, creating schemas, planning integrations, evaluating extraction quality, updating or evolving guidelines, improving system performance, or debugging instruction chains that silently fail, produce inconsistent confidence outputs, loop indefinitely, or write bad data to records.
tools
personize-code
Builds, deploys, and iterates production-ready AI agent pipelines using Trigger.dev and the Personize SDK (code) or n8n (no-code), and authors the advanced multi-step `instructions[]` prompts those pipelines run. Handles the full lifecycle: interview the user about what they want, design the schema and governance, write the pipeline code, deploy it, monitor results, and iterate based on feedback. Generates TypeScript tasks for outbound sequences, inbound lead processing, conversational reply handlers, enrichment pipelines, and account signal monitoring — all backed by Personize memory, AI context, and governance. Also covers connecting external MCP servers, configuring webhook/S3 destinations, and authoring advanced multi-step prompt patterns (conditional branching by account tier, multi-source data reconciliation, compliance-gated generation, persona fanout, bounded research with source triangulation, few-shot calibrated classification, checklist-gated workflows, self-reflective refinement loops). Use this skill whenever someone wants to build an AI agent, automated workflow, email sequence, drip campaign, cold outreach, lead enrichment, reply handler, account monitor, CRM automation, daily digest, or any durable pipeline — whether they provide technical specs or just describe what they want in plain language. Also trigger for Trigger.dev, n8n, background tasks, self-scheduling follow-ups, GTM automation, 'build me an agent that...', 'I want to automate...', MCPs, webhook destinations, or debugging instruction chains that silently fail, produce inconsistent confidence outputs, loop indefinitely, or write bad data to records.
tools