| name | blind-spot-advisor |
| description | Surfaces risks, blind spots, and opportunities the operator didn't ask about. Runs after plans are drafted but before execution. Use when reviewing implementation plans, launching features, or making architectural decisions. |
Blind Spot Advisor
Source: Atlas UX Proactive Advisor pattern, adapted for K3 Agentic Infrastructure
Version: 1.0
Trigger: Run AFTER an implementation plan is drafted, BEFORE execution begins
Purpose
GAG executes what's asked. This skill checks what's NOT being asked. It surfaces blind spots across five domains: compliance, conversion, trust, retention, and competitive positioning.
Execution Protocol
Step 1: Ingest the Plan
Read the implementation plan or feature spec that was just drafted. Identify:
- What is being built
- Who the target user is
- What data flows are involved
- What external services are touched
- What the deployment target is
Step 2: Run the Blind Spot Checklist
Evaluate the plan against EVERY item below. For each category, output findings only if a blind spot exists. Skip categories with no findings.
Blind Spot Checklists
🔴 Compliance & Legal
🟢 Conversion & UX
🔵 Retention & Engagement
⚪ Competitive & Strategic
Output Format
# Blind Spot Advisory — [Plan/Feature Name]
**Reviewed:** [Date]
**Plan Source:** [Link to implementation_plan.md or feature spec]
## 🔴 CRITICAL (Must Address Before Execution)
- [Finding with specific regulatory citation or risk]
## 🟡 HIGH (Should Address Before Launch)
- [Finding with impact assessment]
## 🟢 MEDIUM (Consider for V2)
- [Finding with recommendation]
## 💡 Ideas to Consider
- [Opportunities or enhancements the plan didn't mention]
## ✋ What NOT to Build Yet
- [Features that seem tempting but are premature given current stage]
When to Invoke
| Trigger | Action |
|---|
| Implementation plan drafted | Run full checklist |
| Feature spec written | Run full checklist |
| PR touches auth, payments, or legal content | Run Compliance section only |
| Jules Day Batch task submitted | Run Compliance + Trust sections |
| Quarterly review | Run full checklist against each active feature |
Integration with Action-Gating Loop
This skill maps to Step 2 (VERIFY) and Step 3 (SGL GATE) of the Action-Gating Loop. Findings rated 🔴 CRITICAL should trigger an SGL REVIEW or BLOCK before execution proceeds.