Proof before guidance
The app asks for photos, readings, labels, or confirmations before moving a repair forward.
Safety & trust
AI guidance should slow down at risk.
SkillPlug is designed for real HVAC work, where a confident wrong answer is worse than no answer. The product asks for proof, keeps humans responsible, and stops when work is unsafe.
Safety loop
Proof, then guidance.
This is the safety story the product videos should show: SkillPlug moves only when the evidence is clear.
Photo first
Ask for the unit, label, part, or reading.
Risk pause
Stop before unsafe or license-sensitive work.
Handoff
Leave a useful note for a qualified person.
Live view
Principles
Safety is a product behavior, not a line of copy.
The product needs to prove restraint in the same place it proves intelligence.
Human in control
The worker, instructor, company, or licensed professional remains responsible for the work.
Stop instead of freelancing
Unsafe, uncertain, or out-of-scope work turns into a handoff instead of overconfident advice.
Safety loop
Every guided task follows a bounded loop.
The loop is simple enough for a tech to remember and strict enough for an AI product to respect.
01
Choose repair
The tech or student starts with a narrow repair path.
02
Ask for proof
SkillPlug requests the photo or reading needed for the next step.
03
Stop at risk
Unsafe work, missing proof, or license-sensitive work pauses the flow.
04
Summarize
The session creates a trail for a senior tech, instructor, or company policy.
Mechanisms
What makes the guidance bounded.
These are not decorative trust badges. They are product surfaces that show up during the repair.
Photo requests
Unit photo, nameplate, close-up, reading, or confirmation when the repair path needs proof.
HVAC-specific hazards
Electrical, capacitor, gas, refrigerant, ladder, and rotating-equipment risks in the flow.
Handoff notes
If the job should stop, SkillPlug summarizes symptoms, evidence, likely cause, and what not to do next.
Video story
Show the stop, not just the magic.
For this category, credibility comes from showing where the app refuses to continue.
Safety 01
Ask for a nameplate
The app requests the evidence needed before giving model-specific guidance.
Safety 02
Pause at electrical work
The app stops when the next action requires a qualified person or company procedure.
Safety 03
Create a handoff note
The tech leaves a concise record instead of guessing through the next step.
Questions
Want to discuss the safety model?
We are happy to walk through where SkillPlug guides, where it pauses, and where it hands off to humans.