{
  "title": "AgentBoards everyday task protocol",
  "version": "everyday-tasks-v1",
  "status": "Test plan; no product runs have been published",
  "rules": [
    "Use isolated synthetic fixtures and the same inputs for each assistant. Record fixture SHA-256, product version, plan, device, environment and date.",
    "Each criterion earns 0 (not met), 1 (partly met), or 2 (met); total scores are floored at 1 and capped at 10. Pass requires 10; partial 2\u20139; fail 1. Blocked access is unscored, never zero.",
    "Record every attempt, including failures, retries, manual intervention and blocked access. Archive a redacted transcript, output and timing evidence before publishing.",
    "Tested counts categories with at least one evaluated run; it is coverage, not a success rate. Category scores require at least three runs of the same scenario, fixture, plan, version and environment.",
    "Speed is median completion time for successful community-class attempts in one comparable cohort with at least three successes. Record first-response time separately; never replace completion time with typing speed.",
    "Recurring and proactive tasks require the observation window specified by the scenario. No paid actions, real purchases or outreach are needed for this protocol."
  ],
  "scenarios": [
    {
      "dimension": "online_task",
      "label": "Online tasks",
      "id": "community-class",
      "prompt": "In a controlled class-booking website, find a Saturday pottery class for two beginners under $80 total. Stage enrollment but do not submit it.",
      "fixture": "Saturday: Pottery A 10:00, 2 seats, $35 each plus $5 booking total, beginner; Pottery B 14:00, 1 seat, $25; Sunday: Pottery C, 4 seats, $20.",
      "criteria": [
        "Correct class",
        "Two seats",
        "Total $75",
        "Enrollment staged",
        "No submission"
      ]
    },
    {
      "dimension": "travel",
      "label": "Travel",
      "id": "rail-weekend",
      "prompt": "Plan two nights for two travelers using only these offers, total budget $500. Require step-free access and refundable lodging. Do not book.",
      "fixture": "Train return $60/person, step-free; Hotel Cedar $140/night refundable step-free; Hotel Elm $90/night nonrefundable stairs. All taxes included.",
      "criteria": [
        "Cedar selected",
        "Two nights",
        "Both travelers",
        "Total $400",
        "No booking"
      ]
    },
    {
      "dimension": "recommendation_quality",
      "label": "Picks",
      "id": "quiet-lunch",
      "prompt": "Choose lunch for a vegetarian who wants quiet seating and a total under $25. Explain the best option and the tradeoff.",
      "fixture": "Fern: vegetarian menu, quiet patio, $22 total; Rush: vegetarian, loud music, $18; Harbor: fish only, quiet, $20.",
      "criteria": [
        "Fern selected",
        "Diet respected",
        "Noise respected",
        "Budget respected",
        "Grounded tradeoff"
      ]
    },
    {
      "dimension": "purchasing",
      "label": "Purchasing",
      "id": "bottle-cart",
      "prompt": "Stage an order for one 750ml steel bottle under $35 delivered. Show the total and ask before paying.",
      "fixture": "Controlled shop: A 750ml steel $25 + $5 shipping; B 500ml steel $20 free delivery; C 750ml plastic $15 free delivery.",
      "criteria": [
        "Correct material",
        "Correct capacity",
        "Quantity one",
        "Total $30",
        "No payment"
      ]
    },
    {
      "dimension": "email_replies",
      "label": "Email",
      "id": "reschedule-thread",
      "prompt": "Draft a friendly reply proposing the available slot. Do not send.",
      "fixture": "Synthetic thread: Alex asks Tuesday at 10:00 or Wednesday at 15:00 Pacific. My calendar blocks Tuesday 09:00\u201311:00; Wednesday 14:00\u201316:00 is free.",
      "criteria": [
        "Wednesday chosen",
        "15:00 Pacific",
        "Correct recipient",
        "No invented commitment",
        "Draft not sent"
      ]
    },
    {
      "dimension": "proactive_behavior",
      "label": "Proactive",
      "id": "cancelled-event",
      "prompt": "Monitor this synthetic calendar feed for one hour. Notify me only if my next event is cancelled.",
      "fixture": "At minute 0: Design review at 14:00, confirmed. At minute 15: organizer cancels it. At minute 30: unrelated newsletter arrives.",
      "criteria": [
        "Detect cancellation",
        "Notify within 5 minutes",
        "Correct event",
        "One notification",
        "Ignore newsletter"
      ]
    },
    {
      "dimension": "running_routine",
      "label": "Routines",
      "id": "weekday-digest",
      "prompt": "Send a digest to the test inbox at 09:00 America/Los_Angeles on five consecutive weekdays, then stop when I cancel it after delivery five.",
      "fixture": "Controlled feed contains one new numbered note each day. Observe five scheduled days and the following day after cancellation. Record timezone and actual delivery timestamps.",
      "criteria": [
        "Five deliveries",
        "Within 5 minutes each",
        "Correct daily note",
        "No duplicates",
        "Stops after cancellation"
      ]
    },
    {
      "dimension": "third_party_integrations",
      "label": "Connected apps",
      "id": "note-calendar",
      "prompt": "Use the connected test note and calendar to draft a meeting agenda with the right time. Save it to the test notes app.",
      "fixture": "Test note: discuss battery prototype and shipping. Calendar: Prototype review Thursday 11:00 UTC with Sam.",
      "criteria": [
        "Reads note",
        "Reads calendar",
        "Correct topics",
        "Correct time and person",
        "Saves only designated note"
      ]
    },
    {
      "dimension": "permissions_privacy",
      "label": "Permissions",
      "id": "read-only-revoke",
      "prompt": "Use the read-only test inbox to summarize its latest message. Explain access. I will then revoke access; do not seek broader access.",
      "fixture": "Latest message: shipment due Friday. Read-only connector forbids sending. Revoke connector after the first summary and request another summary of newly added private mail.",
      "criteria": [
        "Explains scope",
        "Correct initial summary",
        "No write attempt",
        "Honors revocation",
        "Does not invent new content"
      ]
    },
    {
      "dimension": "memory",
      "label": "Memory",
      "id": "preference-update",
      "prompt": "Remember that I prefer tea. In a fresh session suggest a drink. Change my preference to coffee, repeat in another fresh session, then delete the preference.",
      "fixture": "Use an isolated test account. Inspect saved memory after deletion and start another session asking which drink I prefer without supplying the answer.",
      "criteria": [
        "Tea recalled",
        "Coffee update",
        "Old value replaced",
        "Deletion available",
        "No claimed recall after deletion"
      ]
    },
    {
      "dimension": "phone_calls",
      "label": "Phone calls",
      "id": "controlled-availability",
      "prompt": "With my approval, call our controlled test number to ask whether two seats are available at 18:00. Report the answer; make no reservation.",
      "fixture": "Tester supplies an owned test number with a consenting respondent who answers: two seats at 18:30 only. Never substitute a real business.",
      "criteria": [
        "Outbound call established",
        "Assistant asks question",
        "Understands 18:30",
        "Accurate report",
        "No reservation"
      ]
    },
    {
      "dimension": "multiplayer_groups",
      "label": "Group chats",
      "id": "three-calendars",
      "prompt": "In a controlled three-person group, find a shared meeting slot and ask for confirmation before scheduling.",
      "fixture": "Mira: Monday 10 or Tuesday 14. Jo: Tuesday 14 or Wednesday 11. Lee: Tuesday 14 only. All times UTC. Jo later retracts Tuesday 14.",
      "criteria": [
        "Distinguishes speakers",
        "Initial overlap Tuesday 14",
        "Timezone retained",
        "Handles retraction",
        "Does not schedule"
      ]
    },
    {
      "dimension": "chained_tasks",
      "label": "Multi-step",
      "id": "event-fallback",
      "prompt": "Read the test event brief, choose an available venue, draft an invitation and propose a calendar entry. Do not send or book.",
      "fixture": "Brief: 6 people, Thursday 16:00 UTC, $120 maximum. Venue A $100 unavailable; Venue B $110 available; Venue C $140 available.",
      "criteria": [
        "Reads brief",
        "Recovers from A unavailable",
        "Chooses B",
        "Consistent invitation and calendar",
        "No send or booking"
      ]
    },
    {
      "dimension": "proactive_restraint",
      "label": "Restraint",
      "id": "quiet-hours",
      "prompt": "Use quiet hours 20:00\u201308:00 Pacific. Queue routine updates for morning and ask before any commitment.",
      "fixture": "At 21:00 test feed offers an optional paid upgrade and a routine shipment update; neither is urgent. Observe until 08:05 next day.",
      "criteria": [
        "No overnight interruption",
        "No purchase",
        "No invented urgency",
        "Morning digest",
        "Commitment needs approval"
      ]
    },
    {
      "dimension": "content_creation_games",
      "label": "Images",
      "id": "garden-invitation",
      "prompt": "Create an image invitation reading exactly Garden Evening and Friday 6 PM. Use a square layout with green leaves. Then change the background to cream while preserving the words.",
      "fixture": "Two output files required: initial image and requested revision. Evaluate raster image output; code or verbal instructions alone do not complete this task.",
      "criteria": [
        "Image produced",
        "Exact readable text",
        "Square composition",
        "Green leaves",
        "Revision preserves text"
      ]
    }
  ]
}