// Comprehensive framework for effective gptme agent onboarding that builds user trust, communicates capabilities clearly, and establishes productive working relationships from the first interaction.
Conduct a structured interview about a new project and generate a ready-to-use CLAUDE.md/AGENTS.md starter so the agent knows your conventions from session one.
Query a Home Assistant instance for presence, sensor data, calendar events, and cameras
Use when creating or revising an agent's SOUL.md, splitting runtime voice out of broader identity docs, or tightening a vague persona file into a short, opinionated, voice-only artifact without changing the agent's core identity.
Workflow for publishing HTML artifacts (demos, visualizations, interactive content) to the web. Enables agents to share work products publicly via GitHub Pages or similar hosting.
Systematic code review workflows with bundled utilities for analyzing code quality, detecting patterns, and providing structured feedback. Use this skill when reviewing pull requests or conducting code audits.
Template and guide for restructuring large documentation files into token-efficient directory structures. Reduces context bloat by 40-60% while maintaining accessibility.
| name | agent-onboarding |
| description | Comprehensive framework for effective gptme agent onboarding that builds user trust, communicates capabilities clearly, and establishes productive working relationships from the first interaction. |
| license | MIT |
| compatibility | Requires gptme |
| metadata | {"author":"bob","version":"1.0.0","tags":["onboarding","trust","user-experience","agent-setup","first-interaction"],"requires_tools":[],"requires_skills":[]} |
A systematic framework for gptme agents to conduct effective user onboarding that maximizes early success and builds long-term trust.
This skill addresses a critical gap in gptme agent deployment: how to transition from technical setup to productive user-agent collaboration. Based on analysis of real agent deployments and user interaction patterns, it provides proven strategies for:
📖 Detailed Reference: For comprehensive implementation details, validation criteria, and advanced patterns, see framework-reference.md.
Apply this skill when:
Before diving into capabilities, assess:
Technical Comfort Level:
Domain Context:
Pace Preference:
High-Tech Professional: "I specialize in [domain] with access to development tools, file analysis, and workflow automation. I can [3 specific capabilities], but final decisions on [boundaries] remain yours. What's your current biggest [domain] challenge?"
Non-Technical Creative: "I'm your project organization assistant. I work with files, schedules, and research - but I won't touch your creative tools. I can help streamline the logistics so you can focus on creating. What part of project management feels overwhelming?"
Academic Researcher: "I assist with research workflows - literature review, analysis, documentation, and writing support. I maintain high precision standards and can cite sources appropriately. I can't replace your expertise, but I can accelerate routine tasks. What research bottleneck should we tackle first?"
Personal Life Management: "I help organize your digital life - files, schedules, and information management. I operate privately and only access what you explicitly share. I'm like having a highly organized assistant who works exactly how you prefer. What area of your life feels most chaotic right now?"
Phase 1 (Interactions 1-3): Demonstrate basic reliability
Phase 2 (Interactions 4-10): Show domain competence
Phase 3 (Interactions 10+): Establish autonomous collaboration
Before First Interaction:
During First Interaction:
Ongoing (Per Session):
1-Week Success Indicators:
1-Month Success Indicators:
Long-Term Success Indicators:
Symptoms: Requests that require reasoning beyond current LLM capabilities, frustration when agent has limitations Recovery: Redirect to specific, demonstrable capabilities. "I excel at [specific domain] tasks like [examples]. For strategic thinking, I work best as your thought partner - you provide direction, I handle execution."
Symptoms: Vague requests, uncertainty about what agent can help with, asks "what can you do?" repeatedly Recovery: Provide specific examples in their domain. "Here are three things I can help with right now: [specific task 1], [specific task 2], [specific task 3]. Which sounds most valuable?"
Symptoms: User requests different level of detail, different formality, different pace Recovery: Adapt immediately and confirm. "I'll adjust to [new style]. Is this level of detail better?"
Symptoms: User hesitant to share context, asks about privacy/security repeatedly, reluctant to try capabilities Recovery: Start with read-only tasks, explain exactly what you're doing, let user approve each step. "I'll only read the file to understand the format - I won't make any changes without your explicit approval."
Symptoms: User stops responding, requests to "slow down," seems confused by multiple options Recovery: Reset to basics. "Let me focus on just one thing: [specific capability]. We can explore other features once this is working smoothly for you."
For comprehensive implementation details, advanced patterns, and validation criteria, see the Framework Reference which includes:
This skill incorporates patterns from:
30-Second User Assessment:
Emergency Recovery Phrases:
If you discover new onboarding patterns or failure modes, contribute them back:
lessons/workflow/agent-onboarding-[scenario].mdThis skill was developed through analysis of real gptme agent deployments and represents synthesized learning from successful and failed onboarding experiences.