/upfront:assess
What it does
Section titled “What it does”/upfront:assess is a focused conversation about a specific concern. Not feature definition — problem-solving. The AI brings domain knowledge you don’t have, maps the landscape, teaches you the risks, and helps you choose between tradeoffs.
When to use it
Section titled “When to use it”- “X seems like a security risk, how should we handle it?”
- “What are the options for caching here?”
- “Is our auth approach solid or are we missing something?”
- “How should we handle rate limiting?”
- Any time you have a specific concern and want to think it through before acting
This skill auto-triggers on phrases like “how should we handle”, “is this a problem”, “what are the options for”, “seems risky.”
How it works
Section titled “How it works”- Understand — restates your concern, asks one clarifying question if needed
- Map the landscape — real risks ranked by likelihood, common approaches, counterintuitive traps
- Explore tradeoffs — explicit option/benefit/cost/failure for each viable approach
- Land on next actions — what to do, what to watch for, what to revisit
Key principle
Section titled “Key principle”This is a teaching conversation. The AI brings knowledge — it doesn’t just ask you questions. You came to it because you don’t know the landscape. It maps it for you, then lets you decide.
Next step
Section titled “Next step”If the outcome is a code change, brainstorm routes you to the right skill: /upfront:quick, /upfront:patch, or /upfront:feature depending on size.