Pattern: Advisory Debate | Team size: 4 agents
This team evaluates each case from multiple expert perspectives before agreeing on a final risk grade and explanation. Advisory debate works well because overstimulation risk is nuanced, involving developmental stage, content design, behavioral signals, and parent communication sensitivity.
The Screen-Time Overstimulation Risk Grader — Parents input their child's age, current favorite apps/games, daily screen duration, and post-screen behaviors (e.g., tantrums, hyperactivity, eye rubbing). The tool outputs an "Overstimulation Risk Score" (Low/Medium/High) with a breakdown of sensory triggers. — It directly targets the primary pain point mentioned in TutoClub's copy: "overstimulation, and guilt for parents."
Create an agent team to design and document the “Screen-Time Overstimulation Risk Grader” for parents, using the Multi-Lens Child Wellbeing Review advisory debate pattern. The tool must let parents input a child’s age, favorite apps/games, daily screen duration, and post-screen behaviors such as tantrums, hyperactivity, eye rubbing, irritability, withdrawal, and sleep disruption. The tool must output an Overstimulation Risk Score of Low, Medium, or High, plus a parent-friendly breakdown of likely sensory triggers. The output must directly address TutoClub’s stated parent pain point: “overstimulation, and guilt for parents,” using supportive, non-shaming language. Create all outputs under this exact project directory: outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/ Use these exact files: 1. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/00_team_brief.md 2. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/01_input_schema.md 3. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/02_developmental_appropriateness_advisor.md 4. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/03_behavioral_signal_advisor.md 5. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/04_digital_product_mechanics_advisor.md 6. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/05_advisory_debate_log.md 7. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/06_consensus_risk_rubric.md 8. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/07_parent_output_templates.md 9. outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/08_synthesis_review.md Follow this workflow exactly. Make dependencies explicit in each file where relevant. Step 1 — Create the team brief. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/00_team_brief.md. Length: 700–900 words. Include these exact sections: # Team Brief ## Business Problem ## Target User ## Product Promise ## Important Safety Boundaries ## Advisory Debate Pattern ## Agent Roles ## Collaboration Rules ## Required Final Outputs Content requirements: - Explain that this is not a medical, psychological, developmental, or vision diagnosis tool. - State that the tool offers educational guidance and should recommend professional support when behaviors are severe, persistent, escalating, or interfering with daily life. - Define the team: - Developmental Appropriateness Advisor: assesses child age, daily duration, timing, and content suitability. - Behavioral Signal Advisor: evaluates post-screen symptoms and their intensity/frequency. - Digital Product Mechanics Advisor: identifies app/game mechanics that may increase stimulation. - Consensus Risk Moderator: resolves disagreements and produces final Low/Medium/High score and parent-facing explanation. - Require agents to challenge each other’s assumptions before final scoring. - Require parent language to be calm, validating, practical, and non-guilt-inducing. This task must complete before Step 2 begins. Step 2 — Define the parent input schema. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/01_input_schema.md. Length: 900–1,100 words. Include these exact sections: # Parent Input Schema ## Required Inputs ## Optional Inputs ## Behavior Checklist ## App/Game Attribute Capture ## Derived Fields ## Missing Data Rules ## Example Completed Input Required inputs must include: - Child age in years and months. - Daily recreational screen duration in minutes. - Most-used apps/games/videos. - Typical timing of use: morning, afternoon, evening, before bed, meals, transitions, car rides, waiting rooms. - Post-screen behaviors within 0–60 minutes after stopping. - Parent concern level: low, medium, high. Behavior checklist must include: - Tantrums/meltdowns - Hyperactivity/restlessness - Irritability - Difficulty transitioning away - Eye rubbing/squinting/headache complaints - Sleep delay/night waking - Withdrawal/flat mood - Aggressive behavior - Craving/asking repeatedly - Sensitivity to noise/light after screen use For each behavior, capture: - Present: yes/no/unsure - Frequency: rarely/sometimes/often/almost always - Intensity: mild/moderate/severe - Duration: under 10 minutes/10–30 minutes/30–60 minutes/over 60 minutes App/game attribute capture must include: - Fast pacing - Bright flashing visuals - Loud or layered audio - Autoplay - Infinite scroll/feed - Variable rewards/loot boxes/random prizes - Competitive pressure/timers - Social pressure/chat - Short-form video - Ads or frequent interruptions Missing data rules must explain how advisors should proceed conservatively without overclaiming. This task must complete before Steps 3, 4, and 5 begin. Step 3 — Run the Developmental Appropriateness Advisor. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/02_developmental_appropriateness_advisor.md. Length: 900–1,100 words. Include these exact sections: # Developmental Appropriateness Advisor Review ## Lens Definition ## Age Bands ## Duration Risk Guidance ## Timing Risk Guidance ## Content Suitability Signals ## Developmental Concerns to Flag ## Scoring Contribution ## Assumptions and Limits ## Questions for Other Advisors Age bands must include: - Under 2 - 2–4 - 5–7 - 8–10 - 11–13 Duration guidance must define low/moderate/high concern for each age band using practical, cautious ranges. Avoid presenting the ranges as medical rules. Use language such as “generally more concerning” and “may increase risk.” Timing guidance must flag: - Screens before bed - Screens during meals - Screens used to manage every transition - Long sessions without breaks - Screens immediately before school or structured activities if behavior worsens afterward Scoring contribution must use a 0–10 advisor score: - 0–3: low developmental concern - 4–6: moderate developmental concern - 7–10: high developmental concern Include at least 5 specific questions this advisor asks the Behavioral Signal Advisor and Digital Product Mechanics Advisor. This task must complete after Step 2 and before Step 6 begins. Step 4 — Run the Behavioral Signal Advisor. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/03_behavioral_signal_advisor.md. Length: 900–1,100 words. Include these exact sections: # Behavioral Signal Advisor Review ## Lens Definition ## Key Post-Screen Signals ## Frequency and Intensity Interpretation ## Red-Flag Patterns ## Protective or Reassuring Patterns ## Scoring Contribution ## Assumptions and Limits ## Questions for Other Advisors Key post-screen signals must analyze: - Tantrums/meltdowns - Hyperactivity/restlessness - Irritability - Difficulty transitioning away - Eye rubbing/squinting/headache complaints - Sleep disruption - Withdrawal/flat mood - Aggression - Repeated craving/asking - Sensory sensitivity after use Red-flag patterns must include: - Severe or almost-always behaviors - Sleep disruption tied to evening use - Aggression or unsafe behavior - Symptoms lasting over 60 minutes - Eye pain, frequent headaches, or persistent visual discomfort - Behavior that interferes with school, relationships, meals, or family routines Scoring contribution must use a 0–10 advisor score: - 0–3: low behavioral concern - 4–6: moderate behavioral concern - 7–10: high behavioral concern Include at least 5 specific questions this advisor asks the Developmental Appropriateness Advisor and Digital Product Mechanics Advisor. This task must complete after Step 2 and before Step 6 begins. Step 5 — Run the Digital Product Mechanics Advisor. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/04_digital_product_mechanics_advisor.md. Length: 900–1,100 words. Include these exact sections: # Digital Product Mechanics Advisor Review ## Lens Definition ## Stimulation Mechanics ## Engagement Loop Mechanics ## Social and Competitive Pressure ## Sensory Load Signals ## App/Game Category Heuristics ## Scoring Contribution ## Assumptions and Limits ## Questions for Other Advisors Stimulation mechanics must cover: - Fast pacing - Bright flashing visuals - Rapid scene changes - Loud or layered audio - Surprise effects - Haptic feedback if relevant Engagement loop mechanics must cover: - Infinite scroll - Autoplay - Variable rewards - Streaks - Loot boxes/random rewards - Push notifications - Timers - Scarcity events - Ads and interruptions App/game category heuristics must include: - Short-form video apps - Sandbox creative games - Competitive multiplayer games - Educational apps - Puzzle games - Streaming cartoons/videos - Social platforms - Casual reward-heavy mobile games Scoring contribution must use a 0–10 advisor score: - 0–3: low mechanics concern - 4–6: moderate mechanics concern - 7–10: high mechanics concern Include at least 5 specific questions this advisor asks the Developmental Appropriateness Advisor and Behavioral Signal Advisor. This task must complete after Step 2 and before Step 6 begins. Step 6 — Conduct the advisory debate. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/05_advisory_debate_log.md. Length: 1,000–1,300 words. Include these exact sections: # Advisory Debate Log ## Case Evaluation Method ## Advisor Opening Positions ## Cross-Advisor Challenges ## Assumption Checks ## Disagreement Resolution ## Consensus Inputs to Final Rubric This task must complete after Steps 3, 4, and 5 and before Step 7 begins. Use this debate structure: 1. Each advisor summarizes their scoring lens in 120–160 words. 2. Each advisor identifies 2 situations where their lens may overestimate risk. 3. Each advisor identifies 2 situations where their lens may underestimate risk. 4. Each advisor challenges at least one assumption from each other advisor. 5. The Consensus Risk Moderator identifies which evidence should receive the most weight. The debate must explicitly resolve these potential conflicts: - High screen duration but no visible behavior problems. - Low screen duration but severe post-screen meltdowns. - Educational app with overstimulating mechanics. - Age-appropriate content used at an inappropriate time, especially before bed. - Parent guilt or anxiety causing over-reporting of concern. - Missing or uncertain app mechanics. Collaboration mechanics: - Advisors must share insights rather than produce isolated opinions. - Advisors must challenge assumptions respectfully and specifically. - Advisors must distinguish evidence from speculation. - Advisors must use uncertainty language when data is incomplete. - Advisors must not shame parents for using screens. Step 7 — Create the consensus risk rubric. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/06_consensus_risk_rubric.md. Length: 1,200–1,500 words. Include these exact sections: # Consensus Risk Rubric ## Overview ## Advisor Score Inputs ## Weighting Model ## Risk Level Definitions ## Sensory Trigger Breakdown ## Override Rules ## Confidence Rating ## Parent-Safe Recommendation Logic ## Worked Examples This task must complete after Step 6 and before Step 8 begins. The weighting model must combine advisor scores as follows: - Developmental Appropriateness Advisor: 30% - Behavioral Signal Advisor: 45% - Digital Product Mechanics Advisor: 25% Define weighted score thresholds: - Low Risk: 0.0–3.4 - Medium Risk: 3.5–6.7 - High Risk: 6.8–10.0 Include override rules: - If severe aggression, unsafe behavior, or symptoms lasting over 60 minutes occur often or almost always, risk cannot be Low. - If sleep disruption is frequent and screen use occurs within 60 minutes of bedtime, risk must be at least Medium. - If eye pain, frequent headaches, or persistent visual discomfort are reported, include a recommendation to consult an appropriate professional. - If the child is under 2 and recreational daily screen use is high, risk must be at least Medium. - If data is missing for app mechanics, do not assume high mechanics risk; mark confidence lower and ask follow-up questions. - If behaviors are mild, rare, and resolve quickly, avoid High unless duration and mechanics are both strongly concerning. Sensory trigger breakdown must categorize triggers as: - Visual intensity - Audio intensity - Reward unpredictability - Transition difficulty - Social/competitive pressure - Timing/sleep vulnerability - Duration/load accumulation Confidence rating must be: - High: most key inputs are present and behavior patterns are clear. - Medium: some uncertainty exists but enough data supports the grade. - Low: multiple important inputs are missing or inconsistent. Worked examples must include exactly 3 examples: 1. Low risk case 2. Medium risk case 3. High risk case Each worked example must include: - Parent input summary: 80–120 words - Advisor scores - Weighted score calculation - Final risk grade - Sensory triggers - Parent-facing explanation: 100–140 words Step 8 — Create parent-facing output templates. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/07_parent_output_templates.md. Length: 1,000–1,300 words. Include these exact sections: # Parent Output Templates ## Tone Rules ## Universal Disclaimer ## Low Risk Template ## Medium Risk Template ## High Risk Template ## Sensory Trigger Explanation Snippets ## Suggested Next-Step Library ## Follow-Up Questions for Missing Data This task must complete after Step 7 and before Step 9 begins. Tone rules: - Use warm, plain language. - Validate that screens are part of modern parenting. - Avoid blame, shame, fear, or moralizing. - Avoid diagnosing the child. - Use “may,” “could,” and “might” for uncertain interpretations. - Emphasize small practical adjustments, not perfection. Universal disclaimer must be 50–80 words and state that the tool is educational, not a diagnosis or medical advice. Each risk template must include: - Risk label - 2–3 sentence summary - 3–5 likely sensory triggers - 3 practical next steps - 1 sentence explaining when to seek professional guidance Suggested next-step library must include at least 12 options, covering: - Shorter sessions - Breaks - Slower-paced content - Reducing autoplay/infinite scroll - Turning off notifications - Avoiding screens before bed - Transition warnings - Co-viewing - Outdoor or movement reset - Audio/brightness adjustments - Replacing reward-heavy apps - Tracking behavior for one week Step 9 — Final synthesis and review. Write outputs/agent_teams_demo/screen_time_overstimulation_risk_grader/08_synthesis_review.md. Length: 900–1,100 words. Include these exact sections: # Synthesis Review ## What the Team Produced ## Final Recommended Risk-Grading Flow ## Key Consensus Decisions ## Parent Trust and Guilt-Reduction Review ## Safety and Scope Review ## Implementation-Ready Checklist ## Open Questions for Product Testing ## Final Recommendation This task must complete after Step 8. The synthesis must: - Summarize how the four agents collaborated. - State the final risk-grading flow in 7–10 numbered steps. - Confirm that the final tool uses Low, Medium, and High risk grades only. - Confirm that the tool provides a sensory trigger breakdown. - Confirm that the tool does not diagnose medical, behavioral, or developmental conditions. - Confirm that parent-facing outputs avoid guilt and shame. - Include an implementation-ready checklist with exactly 10 checklist items. - Include exactly 8 open questions for product testing. - End with a concise final recommendation of 120–160 words. Before finishing, review all files for consistency: - Verify every required file exists at the exact specified path. - Verify all section headings match the required names. - Verify dependency order was followed. - Verify advisor score scales are consistent. - Verify weighted scoring math is consistent. - Verify safety boundaries are repeated where needed. - Verify parent-facing language is supportive and non-diagnostic. - Verify the synthesis/review step is complete.
Create your own AI agent team at Build Agents Store. Describe your business problem and get specialized agent teams with ready-to-use prompts for Claude Code.