{
  "titles": [
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "🗃️ Test Backup Restore in Disposable Dir",
      "first_two_words": [
        "Test",
        "Backup"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t001"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "📢 Notifications — summarize feedback",
      "first_two_words": [
        "Notifications",
        "summarize"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t002"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "✨ Screen-Reader Labels Customer Praise Summary",
      "first_two_words": [
        "Screen-Reader",
        "Labels"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t003"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "⬆️ Pydantic patch — upgrade and test",
      "first_two_words": [
        "Pydantic",
        "patch"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t004"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "🐳 Docker usage — explain inspection",
      "first_two_words": [
        "Docker",
        "usage"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t005"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗃️ Compare SQLite vs PostgreSQL for Offline App",
      "first_two_words": [
        "Compare",
        "SQLite"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t006"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "🗃️ Redis eviction — explain allkeys-lru policy",
      "first_two_words": [
        "Redis",
        "eviction"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t007"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade Pydantic and Run Regression Tests",
      "first_two_words": [
        "Upgrade",
        "Pydantic"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t008"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🚀 Deployment Reverted - Update Incident Timeline",
      "first_two_words": [
        "Deployment",
        "Reverted"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t009"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "📉 Incident timeline — update details",
      "first_two_words": [
        "Incident",
        "timeline"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t010"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "🐳 Docker storage — explain inspection",
      "first_two_words": [
        "Docker",
        "storage"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t011"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade Pydantic and Run Tests",
      "first_two_words": [
        "Upgrade",
        "Pydantic"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t012"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade urllib3 and Run Tests",
      "first_two_words": [
        "Upgrade",
        "urllib3"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t013"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "📝 Summarize Praise for Offline Lessons",
      "first_two_words": [
        "Summarize",
        "Praise"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t014"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗄️ Database Choice — compare options",
      "first_two_words": [
        "Database",
        "Choice"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t015"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "👏 Screen-reader labels — summarize customer praise",
      "first_two_words": [
        "Screen-reader",
        "labels"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t016"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "🐍 Pydantic Version — upgrade and test",
      "first_two_words": [
        "Pydantic",
        "Version"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t017"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🔍 OAuth callbacks — investigate HTTP 400 cause",
      "first_two_words": [
        "OAuth",
        "callbacks"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t018"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "🦀 Rust builds — explain incremental cache misses",
      "first_two_words": [
        "Rust",
        "builds"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t019"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "🐳 Docker disk usage — explain safe inspection",
      "first_two_words": [
        "Docker",
        "disk"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t020"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🔄 Deployment revert — update timeline",
      "first_two_words": [
        "Deployment",
        "revert"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t021"
    },
    {
      "case_id": "feedback-mixed",
      "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
      "title": "🐛 Chess Lesson Feedback: Puzzles Good, Checkmat...",
      "first_two_words": [
        "Chess",
        "Lesson"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t022"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🔒 TLS Certificate Rotation & Webhook Failures",
      "first_two_words": [
        "TLS",
        "Certificate"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t023"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "🗄️ Backup restore — explain safe disposable tes...",
      "first_two_words": [
        "Backup",
        "restore"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t024"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "📊 Database Choice — compare options",
      "first_two_words": [
        "Database",
        "Choice"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t025"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "💄 Customer Praise for New Screen-Reader Labels",
      "first_two_words": [
        "Customer",
        "Praise"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t026"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "🔔 Notification feedback — summarize mixed views",
      "first_two_words": [
        "Notification",
        "feedback"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t027"
    },
    {
      "case_id": "upgrade-pillow",
      "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade Pillow & Run Regression Tests",
      "first_two_words": [
        "Upgrade",
        "Pillow"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t028"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "✨ Notification Scheduling Praised, Default Soun...",
      "first_two_words": [
        "Notification",
        "Scheduling"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t029"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "🗄️ Backup restore — explain safe testing in a d...",
      "first_two_words": [
        "Backup",
        "restore"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t030"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "👷 Docker Disk Usage Inspection Guide",
      "first_two_words": [
        "Docker",
        "Disk"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t031"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "⚠️ Energy Limits — summarize complaints",
      "first_two_words": [
        "Energy",
        "Limits"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t032"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "🧠 Redis Eviction — explain allkeys-lru policy",
      "first_two_words": [
        "Redis",
        "Eviction"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t033"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "📢 Screen-reader labels — summarize praise",
      "first_two_words": [
        "Screen-reader",
        "labels"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t034"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "🐳 Docker disk usage — explain inspection",
      "first_two_words": [
        "Docker",
        "disk"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t035"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "🛠️ Login fix — write release note",
      "first_two_words": [
        "Login",
        "fix"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t036"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "📝 Customer Feedback Summary: Features & Issues",
      "first_two_words": [
        "Customer",
        "Feedback"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t037"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗄️ Database choice — compare options",
      "first_two_words": [
        "Database",
        "choice"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t038"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🔐 TLS rotation — investigate",
      "first_two_words": [
        "TLS",
        "rotation"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t039"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "✅ Testing Backup Restore Safely",
      "first_two_words": [
        "Testing",
        "Backup"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t040"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "😤 Lesson Energy — summarize complaints",
      "first_two_words": [
        "Lesson",
        "Energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t041"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "⬆️ urllib3 — upgrade and test patch release",
      "first_two_words": [
        "urllib3",
        "upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t042"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "😤 Lesson energy limit — summarize complaints",
      "first_two_words": [
        "Lesson",
        "energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t043"
    },
    {
      "case_id": "suspected-leak",
      "request": "Check whether retained image buffers might cause the memory growth. This is only a hypothesis; do not report a confirmed leak.",
      "title": "⚡️ Investigate retained image buffers and memor...",
      "first_two_words": [
        "Investigate",
        "retained"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t044"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🔍 Session timeout — investigate cause",
      "first_two_words": [
        "Session",
        "timeout"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t045"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "😟 Refund delays — summarize complaints",
      "first_two_words": [
        "Refund",
        "delays"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t046"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗃️ SQLite vs PostgreSQL for Offline-First Apps",
      "first_two_words": [
        "SQLite",
        "vs"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t047"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "🔧 Jenkins replacement — draft proposal",
      "first_two_words": [
        "Jenkins",
        "replacement"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t048"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade Pydantic & Run Regression Tests",
      "first_two_words": [
        "Upgrade",
        "Pydantic"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t049"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "👔 Debugging 5-Minute Session Expiration",
      "first_two_words": [
        "Debugging",
        "5-Minute"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t050"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "⬆️ urllib3 — upgrade and run regression tests",
      "first_two_words": [
        "urllib3",
        "upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t051"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🐛 Investigate Login Failure After Password Reset",
      "first_two_words": [
        "Investigate",
        "Login"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t052"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "📝 Lesson Energy — summarize complaints",
      "first_two_words": [
        "Lesson",
        "Energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t053"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🚀 Investigate TLS Rotation Webhook Failures",
      "first_two_words": [
        "Investigate",
        "TLS"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t054"
    },
    {
      "case_id": "upgrade-pillow",
      "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
      "title": "🖼️ Pillow Library — upgrade and test",
      "first_two_words": [
        "Pillow",
        "Library"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t055"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "🔧 Backup error — investigate permission issue",
      "first_two_words": [
        "Backup",
        "error"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t056"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🐛 Zephyr Unicode CSV Parser Crash Investigation",
      "first_two_words": [
        "Zephyr",
        "Unicode"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t057"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🐛 Investigate TLS rotation and webhook failures",
      "first_two_words": [
        "Investigate",
        "TLS"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t058"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🔑 Login Failure — investigate reset issue",
      "first_two_words": [
        "Login",
        "Failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t059"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "🔐 Login fix — write release note",
      "first_two_words": [
        "Login",
        "fix"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t060"
    },
    {
      "case_id": "suspected-leak",
      "request": "Check whether retained image buffers might cause the memory growth. This is only a hypothesis; do not report a confirmed leak.",
      "title": "🔍 Image buffers — investigate memory growth",
      "first_two_words": [
        "Image",
        "buffers"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t061"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🔍 Session Timeout — investigate cause",
      "first_two_words": [
        "Session",
        "Timeout"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t062"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "🏗️ Buildkite Proposal — draft replacement",
      "first_two_words": [
        "Buildkite",
        "Proposal"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t063"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "🐛 Lesson energy limit interrupts practice and f...",
      "first_two_words": [
        "Lesson",
        "energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t064"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "🐛 Write Release Note for Login Fix",
      "first_two_words": [
        "Write",
        "Release"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t065"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🐛 Investigate OAuth Callback HTTP 400 Errors",
      "first_two_words": [
        "Investigate",
        "OAuth"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t066"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "🚆 Offline lessons — summarize customer praise",
      "first_two_words": [
        "Offline",
        "lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t067"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🔍 Session expiry — investigate 5-minute timeout",
      "first_two_words": [
        "Session",
        "expiry"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t068"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "📝 Inspect Docker Disk Usage Safely",
      "first_two_words": [
        "Inspect",
        "Docker"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t069"
    },
    {
      "case_id": "feedback-mixed",
      "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
      "title": "🐛 Chess Lessons: Puzzle Praise, Checkmate Compl...",
      "first_two_words": [
        "Chess",
        "Lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t070"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🔎 Session expiry — investigate five-minute timeout",
      "first_two_words": [
        "Session",
        "expiry"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t071"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "🏗️ Jenkins migration — draft proposal",
      "first_two_words": [
        "Jenkins",
        "migration"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t072"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "🗃️ Test Database Restore in Disposable Dir",
      "first_two_words": [
        "Test",
        "Database"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t073"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "📝 Summarize Lesson Energy Complaints",
      "first_two_words": [
        "Summarize",
        "Lesson"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t074"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "💾 Backup Restore — explain testing process",
      "first_two_words": [
        "Backup",
        "Restore"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t075"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "📝 Weeks-Long Refund Processing Complaints",
      "first_two_words": [
        "Weeks-Long",
        "Refund"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t076"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "⭐ Offline lessons — summarize praise",
      "first_two_words": [
        "Offline",
        "lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t077"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "🔧 Inspect Docker Disk Usage",
      "first_two_words": [
        "Inspect",
        "Docker"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t078"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗃️ Database choice — compare SQLite and PostgreSQL",
      "first_two_words": [
        "Database",
        "choice"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t079"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "📝 Summarize Customer Refund Delay Complaints",
      "first_two_words": [
        "Summarize",
        "Customer"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t080"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🕒 Incident Timeline — update details",
      "first_two_words": [
        "Incident",
        "Timeline"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t081"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "🐛 Write Login Fix Release Note",
      "first_two_words": [
        "Write",
        "Login"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t082"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🐛 Debugging Login Failure After Password Reset",
      "first_two_words": [
        "Debugging",
        "Login"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t083"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade urllib3 & Run Tests",
      "first_two_words": [
        "Upgrade",
        "urllib3"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t084"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🐛 Investigate Zephyr's Unicode CSV Parser Crash",
      "first_two_words": [
        "Investigate",
        "Zephyr's"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t085"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "📝 Write Login Fix Release Note",
      "first_two_words": [
        "Write",
        "Login"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t086"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "♿ Screen-reader labels — summarize praise",
      "first_two_words": [
        "Screen-reader",
        "labels"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t087"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "🐛 Investigate nightly backup permission error",
      "first_two_words": [
        "Investigate",
        "nightly"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t088"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🚀 Update Incident Timeline for Reverted Deployment",
      "first_two_words": [
        "Update",
        "Incident"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t089"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🔍 OAuth callbacks — investigate 400 error",
      "first_two_words": [
        "OAuth",
        "callbacks"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t090"
    },
    {
      "case_id": "feedback-mixed",
      "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
      "title": "🐛 Mixed Chess Feedback: Puzzles and Checkmate Rule",
      "first_two_words": [
        "Mixed",
        "Chess"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t091"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🔍 TLS certificate — investigate",
      "first_two_words": [
        "TLS",
        "certificate"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t092"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "📝 Summarize Customer Praise for Screen-Reader L...",
      "first_two_words": [
        "Summarize",
        "Customer"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t093"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "🦀 Rust Compilation — explain cache misses",
      "first_two_words": [
        "Rust",
        "Compilation"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t094"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "👷 Buildkite Replacement Proposal for Jenkins",
      "first_two_words": [
        "Buildkite",
        "Replacement"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t095"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "👔 Customer Refund Processing Delays",
      "first_two_words": [
        "Customer",
        "Refund"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t096"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "⚡️ Rust Incremental Compilation & Cache Misses",
      "first_two_words": [
        "Rust",
        "Incremental"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t097"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "✅ Test Backup Restore in Disposable Dir",
      "first_two_words": [
        "Test",
        "Backup"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t098"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "👷 Rust Incremental Compilation Cache",
      "first_two_words": [
        "Rust",
        "Incremental"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t099"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🐛 Investigate TLS Rotation and Webhook Failures",
      "first_two_words": [
        "Investigate",
        "TLS"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t100"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "✅ Login Fix — write release note",
      "first_two_words": [
        "Login",
        "Fix"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t101"
    },
    {
      "case_id": "suspected-leak",
      "request": "Check whether retained image buffers might cause the memory growth. This is only a hypothesis; do not report a confirmed leak.",
      "title": "🔍 Image buffers — investigate memory growth hyp...",
      "first_two_words": [
        "Image",
        "buffers"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t102"
    },
    {
      "case_id": "upgrade-pillow",
      "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade Pillow and Run Regression Tests",
      "first_two_words": [
        "Upgrade",
        "Pillow"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t103"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🚀 Update deployment incident timeline",
      "first_two_words": [
        "Update",
        "deployment"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t104"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "⚡️ Redis Cache Eviction & allkeys-lru Policy",
      "first_two_words": [
        "Redis",
        "Cache"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t105"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "📝 Explain How to Inspect Docker Disk Usage",
      "first_two_words": [
        "Explain",
        "How"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t106"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗄️ SQLite vs PostgreSQL — compare",
      "first_two_words": [
        "SQLite",
        "vs"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t107"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🔑 OAuth callbacks — investigate 400 error",
      "first_two_words": [
        "OAuth",
        "callbacks"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t108"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "📝 Rust Incremental Compilation and Cache Misses",
      "first_two_words": [
        "Rust",
        "Incremental"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t109"
    },
    {
      "case_id": "suspected-leak",
      "request": "Check whether retained image buffers might cause the memory growth. This is only a hypothesis; do not report a confirmed leak.",
      "title": "⚡️ Investigate Retained Image Buffers and Memor...",
      "first_two_words": [
        "Investigate",
        "Retained"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t110"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🔍 Login failure — investigate password reset",
      "first_two_words": [
        "Login",
        "failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t111"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🕵️ Session Expiry — investigate timeout",
      "first_two_words": [
        "Session",
        "Expiry"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t112"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🐞 Zephyr CSV parser — investigate Unicode crash",
      "first_two_words": [
        "Zephyr",
        "CSV"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t113"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "⬆️ Pydantic Version — upgrade/test",
      "first_two_words": [
        "Pydantic",
        "Version"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t114"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "✨ Summarize Offline Lessons Customer Praise",
      "first_two_words": [
        "Summarize",
        "Offline"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t115"
    },
    {
      "case_id": "suspected-leak",
      "request": "Check whether retained image buffers might cause the memory growth. This is only a hypothesis; do not report a confirmed leak.",
      "title": "🖼️ Image Buffers — investigate memory growth",
      "first_two_words": [
        "Image",
        "Buffers"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t116"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗄️ SQLite vs PostgreSQL — compare options",
      "first_two_words": [
        "SQLite",
        "vs"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t117"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "🗃️ Redis Cache Eviction and allkeys-lru Policy",
      "first_two_words": [
        "Redis",
        "Cache"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t118"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "🔍 Backup failure — investigate permission error",
      "first_two_words": [
        "Backup",
        "failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t119"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "✨ Summarize Screen-Reader Label Praise",
      "first_two_words": [
        "Summarize",
        "Screen-Reader"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t120"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🔐 Login failure — investigate reset issue",
      "first_two_words": [
        "Login",
        "failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t121"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "👷 Proposal to Replace Jenkins with Buildkite",
      "first_two_words": [
        "Proposal",
        "to"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t122"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "📝 Login Fix — write release note",
      "first_two_words": [
        "Login",
        "Fix"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t123"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "🔍 Backup permissions — investigate failure",
      "first_two_words": [
        "Backup",
        "permissions"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t124"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "👥 Offline lessons — summarize praise",
      "first_two_words": [
        "Offline",
        "lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t125"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "📝 Notification Scheduling Feedback Summary",
      "first_two_words": [
        "Notification",
        "Scheduling"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t126"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "👷 Jenkins to Buildkite Migration Proposal",
      "first_two_words": [
        "Jenkins",
        "to"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t127"
    },
    {
      "case_id": "feedback-mixed",
      "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
      "title": "♟️ Chess Lessons — summarize mixed feedback",
      "first_two_words": [
        "Chess",
        "Lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t128"
    },
    {
      "case_id": "upgrade-pillow",
      "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
      "title": "🖼️ Pillow Upgrade — upgrade/test",
      "first_two_words": [
        "Pillow",
        "Upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t129"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🔎 OAuth callbacks — investigate HTTP 400 cause",
      "first_two_words": [
        "OAuth",
        "callbacks"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t130"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "🔔 Notification feedback — summarize",
      "first_two_words": [
        "Notification",
        "feedback"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t131"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade urllib3 and run regression tests",
      "first_two_words": [
        "Upgrade",
        "urllib3"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t132"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "🛠️ Jenkins replacement — draft proposal",
      "first_two_words": [
        "Jenkins",
        "replacement"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t133"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🔐 Login failure — investigate after reset",
      "first_two_words": [
        "Login",
        "failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t134"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🔍 Zephyr Parser — investigate Unicode crash",
      "first_two_words": [
        "Zephyr",
        "Parser"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t135"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "📉 Refund delays — summarize complaints",
      "first_two_words": [
        "Refund",
        "delays"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t136"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🌐 Webhook Failures — investigate TLS rotation",
      "first_two_words": [
        "Webhook",
        "Failures"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t137"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "📦 Upgrade Pydantic & Run Regression Tests",
      "first_two_words": [
        "Upgrade",
        "Pydantic"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t138"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "📝 Summarize Lesson Energy Limit Complaints",
      "first_two_words": [
        "Summarize",
        "Lesson"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t139"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "📦 Pydantic — upgrade and test",
      "first_two_words": [
        "Pydantic",
        "upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t140"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "⚡ Rust Incremental Compilation & Cache Misses",
      "first_two_words": [
        "Rust",
        "Incremental"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t141"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🔒️ Investigate 5-Minute Session Expiration",
      "first_two_words": [
        "Investigate",
        "5-Minute"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t142"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "🔧 urllib3 — upgrade and test",
      "first_two_words": [
        "urllib3",
        "upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t143"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🐛 Fix Zephyr Unicode CSV Parser Crash",
      "first_two_words": [
        "Fix",
        "Zephyr"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t144"
    },
    {
      "case_id": "upgrade-pillow",
      "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
      "title": "📦 Pillow library — upgrade and test",
      "first_two_words": [
        "Pillow",
        "library"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t145"
    },
    {
      "case_id": "upgrade-pillow",
      "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
      "title": "🖼️ Pillow — upgrade and test regression suite",
      "first_two_words": [
        "Pillow",
        "upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t146"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🚀 Update incident timeline after rollback",
      "first_two_words": [
        "Update",
        "incident"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t147"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "🦀 Rust incremental compilation — explain cache ...",
      "first_two_words": [
        "Rust",
        "incremental"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t148"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "📦️ Upgrade urllib3 and Run Regression Tests",
      "first_two_words": [
        "Upgrade",
        "urllib3"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t149"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗃️ SQLite vs. PostgreSQL for Offline-First Apps",
      "first_two_words": [
        "SQLite",
        "vs"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t150"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "🐛 Lesson Energy Limit Complaint Summary",
      "first_two_words": [
        "Lesson",
        "Energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t151"
    },
    {
      "case_id": "filler-build",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Rust incremental compilation and why cache misses slow builds.",
      "title": "🦀 Rust compilation — explain cache misses",
      "first_two_words": [
        "Rust",
        "compilation"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t152"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🕒 Deployment Incident — update timeline",
      "first_two_words": [
        "Deployment",
        "Incident"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t153"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🔒️ Session Expiration Investigation",
      "first_two_words": [
        "Session",
        "Expiration"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t154"
    },
    {
      "case_id": "suspected-leak",
      "request": "Check whether retained image buffers might cause the memory growth. This is only a hypothesis; do not report a confirmed leak.",
      "title": "🖼️ Image buffers — investigate memory growth hy...",
      "first_two_words": [
        "Image",
        "buffers"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t155"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "✨ Customer Praise for Screen-Reader Labels",
      "first_two_words": [
        "Customer",
        "Praise"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t156"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "🐛 Lesson Energy Limit Practice Interruption",
      "first_two_words": [
        "Lesson",
        "Energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t157"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "🗃️ Test a Backup Restore Safely in a Disposable...",
      "first_two_words": [
        "Test",
        "a"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t158"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🔍 Zephyr CSV — investigate crash",
      "first_two_words": [
        "Zephyr",
        "CSV"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t159"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "🔍 Nightly backup — investigate permission error",
      "first_two_words": [
        "Nightly",
        "backup"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t160"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "📝 Summarize Screen-Reader Label Praise",
      "first_two_words": [
        "Summarize",
        "Screen-Reader"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t161"
    },
    {
      "case_id": "negation",
      "request": "Explain how to inspect Docker disk usage. Do not delete containers, images, or volumes.",
      "title": "👷 Inspect Docker Disk Usage Guide",
      "first_two_words": [
        "Inspect",
        "Docker"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t162"
    },
    {
      "case_id": "suspected-leak",
      "request": "Check whether retained image buffers might cause the memory growth. This is only a hypothesis; do not report a confirmed leak.",
      "title": "⚡️ Investigate Image Buffer Memory Growth",
      "first_two_words": [
        "Investigate",
        "Image"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t163"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "🐍 urllib3 Upgrade — upgrade/test",
      "first_two_words": [
        "urllib3",
        "Upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t164"
    },
    {
      "case_id": "upgrade-pydantic",
      "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
      "title": "⬆️ Pydantic — upgrade and test regression suite",
      "first_two_words": [
        "Pydantic",
        "upgrade"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t165"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "📝 Summarize Refund Delay Complaints",
      "first_two_words": [
        "Summarize",
        "Refund"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t166"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🔍 Login failure — investigate reset issue",
      "first_two_words": [
        "Login",
        "failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t167"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "🗄️ Redis eviction — explain allkeys-lru",
      "first_two_words": [
        "Redis",
        "eviction"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t168"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🐛 Zephyr CSV parser — investigate Unicode crash",
      "first_two_words": [
        "Zephyr",
        "CSV"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t169"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🐛 Debug OAuth 400 Callback Errors",
      "first_two_words": [
        "Debug",
        "OAuth"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t170"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🔒️ TLS Certificate Rotation & Webhook Failures",
      "first_two_words": [
        "TLS",
        "Certificate"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t171"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🕒 Incident timeline — update rollback record",
      "first_two_words": [
        "Incident",
        "timeline"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t172"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🐛 OAuth 400 Error Investigation",
      "first_two_words": [
        "OAuth",
        "400"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t173"
    },
    {
      "case_id": "uncertainty",
      "request": "Check whether TLS certificate rotation might explain the intermittent webhook failures. This is a hypothesis, not a confirmed cause.",
      "title": "🔐 Webhook TLS — investigate rotation hypothesis",
      "first_two_words": [
        "Webhook",
        "TLS"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t174"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🔑 OAuth Callbacks — investigate 400 error",
      "first_two_words": [
        "OAuth",
        "Callbacks"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t175"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "🔧 Debug Nightly Backup Permission Error",
      "first_two_words": [
        "Debug",
        "Nightly"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t176"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "🔔 Notification Feedback — summarize mixed views",
      "first_two_words": [
        "Notification",
        "Feedback"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t177"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "👔 Debugging Session Expiration Timeout",
      "first_two_words": [
        "Debugging",
        "Session"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t178"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "📚 Redis eviction — explain allkeys-lru",
      "first_two_words": [
        "Redis",
        "eviction"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t179"
    },
    {
      "case_id": "feedback-mixed",
      "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
      "title": "🐛 Fix Chess Checkmate Logic & Summarize Feedback",
      "first_two_words": [
        "Fix",
        "Chess"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t180"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "📝 Test a Backup Restore in a Disposable Directory",
      "first_two_words": [
        "Test",
        "a"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t181"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "📝 Summarize Notification Feedback",
      "first_two_words": [
        "Summarize",
        "Notification"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t182"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗃️ SQLite vs PostgreSQL for Offline App",
      "first_two_words": [
        "SQLite",
        "vs"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t183"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "🐛 Lesson energy limit interrupts practice",
      "first_two_words": [
        "Lesson",
        "energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t184"
    },
    {
      "case_id": "auth-session",
      "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
      "title": "🐛 Investigate Five-Minute Session Expiration",
      "first_two_words": [
        "Investigate",
        "Five-Minute"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t185"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "👔 Refund Processing Delays Report",
      "first_two_words": [
        "Refund",
        "Processing"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t186"
    },
    {
      "case_id": "feedback-notifications",
      "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
      "title": "📝 Customer Feedback: Notifications Analysis",
      "first_two_words": [
        "Customer",
        "Feedback"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t187"
    },
    {
      "case_id": "feedback-mixed",
      "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
      "title": "📝 Summarize Chess Lesson Feedback",
      "first_two_words": [
        "Summarize",
        "Chess"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t188"
    },
    {
      "case_id": "upgrade-urllib3",
      "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
      "title": "📦 urllib3 patch — upgrade and test",
      "first_two_words": [
        "urllib3",
        "patch"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t189"
    },
    {
      "case_id": "docs-after-revert",
      "request": "The deployment was reverted yesterday. Update the incident timeline; do not redeploy or reopen the rollout.",
      "title": "🔄 Deployment Revert — update timeline",
      "first_two_words": [
        "Deployment",
        "Revert"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t190"
    },
    {
      "case_id": "feedback-complaint",
      "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
      "title": "⏳ Lesson energy — summarize complaints",
      "first_two_words": [
        "Lesson",
        "energy"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t191"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🔒️ Login Failure After Password Reset Investiga...",
      "first_two_words": [
        "Login",
        "Failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t192"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "👥 Screen-reader labels — summarize feedback",
      "first_two_words": [
        "Screen-reader",
        "labels"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t193"
    },
    {
      "case_id": "feedback-accessibility",
      "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
      "title": "👥 Screen-reader labels — summarize praise",
      "first_two_words": [
        "Screen-reader",
        "labels"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t194"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "📝 Summarize Offline Lessons Praise",
      "first_two_words": [
        "Summarize",
        "Offline"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t195"
    },
    {
      "case_id": "completed",
      "request": "The login fix has already merged and all regression tests passed. Write a short release note about the completed fix.",
      "title": "🐛 Login Fix Release Note",
      "first_two_words": [
        "Login",
        "Fix"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t196"
    },
    {
      "case_id": "proposal-not-approved",
      "request": "Draft a proposal to replace Jenkins with Buildkite. No migration has been approved or started.",
      "title": "🔨 Jenkins replacement — draft proposal",
      "first_two_words": [
        "Jenkins",
        "replacement"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t197"
    },
    {
      "case_id": "feedback-mixed",
      "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
      "title": "♟️ Chess lessons — summarize mixed feedback",
      "first_two_words": [
        "Chess",
        "lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t198"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "✨ Offline Lessons — summarize praise",
      "first_two_words": [
        "Offline",
        "Lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t199"
    },
    {
      "case_id": "backup-restore",
      "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
      "title": "🔄 Backup restore — test safely",
      "first_two_words": [
        "Backup",
        "restore"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t200"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "😤 Refund delays — summarize complaints",
      "first_two_words": [
        "Refund",
        "delays"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t201"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "🚆 Offline Lessons — summarize praise",
      "first_two_words": [
        "Offline",
        "Lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t202"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "📋 Refund delays — summarize complaints",
      "first_two_words": [
        "Refund",
        "delays"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t203"
    },
    {
      "case_id": "feedback-praise",
      "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
      "title": "✨ Offline Lessons Let Users Practice on the Train",
      "first_two_words": [
        "Offline",
        "Lessons"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t204"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "📝 Explaining Redis Cache Eviction and allkeys-lru",
      "first_two_words": [
        "Explaining",
        "Redis"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t205"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "😟 Refund delays — summarize customer frustration",
      "first_two_words": [
        "Refund",
        "delays"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t206"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "📝 Redis Cache Eviction and allkeys-lru Policy",
      "first_two_words": [
        "Redis",
        "Cache"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t207"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "📝 Summarize Frustration Over Weeks-Long Refunds",
      "first_two_words": [
        "Summarize",
        "Frustration"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t208"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "👷 Investigate Nightly Backup Permission Error",
      "first_two_words": [
        "Investigate",
        "Nightly"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t209"
    },
    {
      "case_id": "neutral-compare",
      "request": "Compare SQLite and PostgreSQL for a small offline-first app. I have not selected either database.",
      "title": "🗃️ SQLite vs PostgreSQL for Offline-First App",
      "first_two_words": [
        "SQLite",
        "vs"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t210"
    },
    {
      "case_id": "auth-login",
      "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
      "title": "🔒️ Login Failure After Password Reset",
      "first_two_words": [
        "Login",
        "Failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t211"
    },
    {
      "case_id": "feedback-refunds",
      "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
      "title": "⚠️ Refund Delays — summarize complaints",
      "first_two_words": [
        "Refund",
        "Delays"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity",
        "connotation"
      ],
      "id": "t212"
    },
    {
      "case_id": "auth-oauth",
      "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
      "title": "🐛 Debugging OAuth Callback HTTP 400 Errors",
      "first_two_words": [
        "Debugging",
        "OAuth"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t213"
    },
    {
      "case_id": "long-tail",
      "request": "Reference background only; the actual request follows at the end.\nThe project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. The project uses a small command-line interface and stores its examples in a dedicated folder. \nACTUAL REQUEST: Investigate Zephyr's Unicode CSV parser crash. Do not work on the CLI examples.",
      "title": "🔍 Zephyr parser — investigate crash",
      "first_two_words": [
        "Zephyr",
        "parser"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t214"
    },
    {
      "case_id": "filler-cache",
      "request": "I was hoping you could help me with a task today. What I need is an explanation of Redis cache eviction, especially the allkeys-lru policy.",
      "title": "⚡️ Redis Cache Eviction & allkeys-lru",
      "first_two_words": [
        "Redis",
        "Cache"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t215"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "👷 Investigate nightly backup permission error",
      "first_two_words": [
        "Investigate",
        "nightly"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t216"
    },
    {
      "case_id": "backup-status",
      "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
      "title": "💾 Backup Failure — investigate error",
      "first_two_words": [
        "Backup",
        "Failure"
      ],
      "criteria": [
        "clarity",
        "frontload",
        "fidelity"
      ],
      "id": "t217"
    }
  ],
  "groups": [
    {
      "id": "g016",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "🔍 Backup failure — investigate permission error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "🔄 Backup restore — test safely"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g001",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔒️ Login Failure After Password Reset"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🔒️ Investigate 5-Minute Session Expiration"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🐛 OAuth 400 Error Investigation"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g047",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "⏳ Lesson energy — summarize complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "🚆 Offline lessons — summarize customer praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "♟️ Chess lessons — summarize mixed feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "😟 Refund delays — summarize customer frustration"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "♿ Screen-reader labels — summarize praise"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "🔔 Notification feedback — summarize mixed views"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g040",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "🐛 Investigate nightly backup permission error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "🗃️ Test a Backup Restore Safely in a Disposable..."
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g043",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "📝 Lesson Energy — summarize complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "🚆 Offline lessons — summarize customer praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "♟️ Chess lessons — summarize mixed feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "😟 Refund delays — summarize complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "👏 Screen-reader labels — summarize customer praise"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "🔔 Notification feedback — summarize mixed views"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g004",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "👷 Investigate Nightly Backup Permission Error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "✅ Test Backup Restore in Disposable Dir"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g034",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade urllib3 and run regression tests"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pydantic and Run Regression Tests"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pillow and Run Regression Tests"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g048",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "🔍 Nightly backup — investigate permission error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "🗄️ Backup restore — explain safe disposable tes..."
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g042",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "⬆️ urllib3 — upgrade and run regression tests"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "⬆️ Pydantic — upgrade and test regression suite"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "🖼️ Pillow — upgrade and test regression suite"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g010",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "📦 urllib3 patch — upgrade and test"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦 Pydantic — upgrade and test"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦 Pillow library — upgrade and test"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g015",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "😤 Lesson Energy — summarize complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "⭐ Offline lessons — summarize praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "♟️ Chess lessons — summarize mixed feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "📋 Refund delays — summarize complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "👥 Screen-reader labels — summarize feedback"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "📢 Notifications — summarize feedback"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g020",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "🔧 Debug Nightly Backup Permission Error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "🗃️ Test Backup Restore in Disposable Dir"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g032",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "💾 Backup Failure — investigate error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "💾 Backup Restore — explain testing process"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g008",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "👷 Investigate Nightly Backup Permission Error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "✅ Testing Backup Restore Safely"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g002",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade urllib3 & Run Tests"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦 Upgrade Pydantic & Run Regression Tests"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pillow & Run Regression Tests"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g030",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "🐍 urllib3 Upgrade — upgrade/test"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "⬆️ Pydantic Version — upgrade/test"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "🖼️ Pillow Library — upgrade and test"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g028",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "💾 Backup Failure — investigate error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "💾 Backup Restore — explain testing process"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g044",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "🔍 Backup permissions — investigate failure"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "🗄️ Backup restore — explain safe testing in a d..."
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g009",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔍 Login failure — investigate reset issue"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🔍 Session Timeout — investigate cause"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🔍 OAuth callbacks — investigate 400 error"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g023",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "📝 Summarize Lesson Energy Complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "📝 Summarize Offline Lessons Praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "📝 Summarize Chess Lesson Feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "📝 Summarize Customer Refund Delay Complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "📝 Summarize Screen-Reader Label Praise"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "📝 Summarize Notification Feedback"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g017",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🐛 Debugging Login Failure After Password Reset"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "👔 Debugging 5-Minute Session Expiration"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🐛 Debugging OAuth Callback HTTP 400 Errors"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g039",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "🐛 Lesson energy limit interrupts practice and f..."
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "✨ Offline Lessons Let Users Practice on the Train"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "🐛 Chess Lessons: Puzzle Praise, Checkmate Compl..."
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "📝 Weeks-Long Refund Processing Complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "💄 Customer Praise for New Screen-Reader Labels"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "📝 Notification Scheduling Feedback Summary"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g046",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "⬆️ urllib3 — upgrade and test patch release"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "⬆️ Pydantic patch — upgrade and test"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "🖼️ Pillow — upgrade and test regression suite"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g037",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🐛 Investigate Login Failure After Password Reset"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🐛 Investigate Five-Minute Session Expiration"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🐛 Investigate OAuth Callback HTTP 400 Errors"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g018",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade urllib3 and Run Regression Tests"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pydantic and Run Tests"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pillow and Run Regression Tests"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g045",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔐 Login failure — investigate after reset"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🔎 Session expiry — investigate five-minute timeout"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🔍 OAuth callbacks — investigate HTTP 400 cause"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g024",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "🔧 Debug Nightly Backup Permission Error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "🗃️ Test Database Restore in Disposable Dir"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g013",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔍 Login failure — investigate password reset"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🔍 Session timeout — investigate cause"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🔍 OAuth callbacks — investigate 400 error"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g019",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "📝 Summarize Lesson Energy Limit Complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "📝 Summarize Offline Lessons Praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "🐛 Fix Chess Checkmate Logic & Summarize Feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "📝 Summarize Refund Delay Complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "✨ Summarize Screen-Reader Label Praise"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "📝 Summarize Notification Feedback"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g005",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔒️ Login Failure After Password Reset Investiga..."
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🔒️ Session Expiration Investigation"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🐛 Debug OAuth 400 Callback Errors"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g033",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🐛 Investigate Login Failure After Password Reset"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🐛 Investigate Five-Minute Session Expiration"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🐛 Investigate OAuth Callback HTTP 400 Errors"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g027",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "⚠️ Energy Limits — summarize complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "✨ Offline Lessons — summarize praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "♟️ Chess Lessons — summarize mixed feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "⚠️ Refund Delays — summarize complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "📢 Screen-reader labels — summarize praise"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "🔔 Notification Feedback — summarize mixed views"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g031",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "⚠️ Energy Limits — summarize complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "🚆 Offline Lessons — summarize praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "♟️ Chess lessons — summarize mixed feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "📉 Refund delays — summarize complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "📢 Screen-reader labels — summarize praise"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "🔔 Notification Feedback — summarize mixed views"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g014",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "🔧 urllib3 — upgrade and test"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦 Pydantic — upgrade and test"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦 Pillow library — upgrade and test"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g041",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔐 Login failure — investigate reset issue"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🔍 Session expiry — investigate 5-minute timeout"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🔎 OAuth callbacks — investigate HTTP 400 cause"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g011",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "😤 Lesson energy limit — summarize complaints"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "👥 Offline lessons — summarize praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "♟️ Chess lessons — summarize mixed feedback"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "😤 Refund delays — summarize complaints"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "👥 Screen-reader labels — summarize praise"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "🔔 Notification feedback — summarize"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g007",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "🐛 Lesson Energy Limit Complaint Summary"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "✨ Summarize Offline Lessons Customer Praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "🐛 Chess Lesson Feedback: Puzzles Good, Checkmat..."
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "👔 Customer Refund Processing Delays"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "✨ Customer Praise for Screen-Reader Labels"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "📝 Customer Feedback Summary: Features & Issues"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g006",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade urllib3 and Run Regression Tests"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pydantic & Run Regression Tests"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pillow & Run Regression Tests"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g012",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "🔧 Backup error — investigate permission issue"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "🔄 Backup restore — test safely"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g021",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🐛 Debugging Login Failure After Password Reset"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "👔 Debugging Session Expiration Timeout"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🐛 Debugging OAuth Callback HTTP 400 Errors"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g025",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔑 Login Failure — investigate reset issue"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🕵️ Session Expiry — investigate timeout"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🔑 OAuth callbacks — investigate 400 error"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g038",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade urllib3 and Run Regression Tests"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pydantic and Run Regression Tests"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pillow and Run Regression Tests"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g035",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "🐛 Lesson energy limit interrupts practice"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "📝 Summarize Praise for Offline Lessons"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "🐛 Mixed Chess Feedback: Puzzles and Checkmate Rule"
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "📝 Summarize Frustration Over Weeks-Long Refunds"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "📝 Summarize Customer Praise for Screen-Reader L..."
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "✨ Notification Scheduling Praised, Default Soun..."
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g029",
      "items": [
        {
          "request": "Investigate why login fails after a password reset. No cause is confirmed yet.",
          "title": "🔑 Login Failure — investigate reset issue"
        },
        {
          "request": "Investigate why sessions expire after five minutes. No cause is confirmed yet.",
          "title": "🕵️ Session Expiry — investigate timeout"
        },
        {
          "request": "Investigate why OAuth callbacks return HTTP 400. No cause is confirmed yet.",
          "title": "🔑 OAuth Callbacks — investigate 400 error"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g003",
      "items": [
        {
          "request": "Summarize customer complaints: users say the lesson energy limit interrupts practice and forces them to wait. Do not turn this into praise.",
          "title": "🐛 Lesson Energy Limit Practice Interruption"
        },
        {
          "request": "Summarize customer praise: users love the offline lessons because they can practice on the train. Do not invent a complaint.",
          "title": "✨ Summarize Offline Lessons Customer Praise"
        },
        {
          "request": "Summarize mixed feedback about chess lessons: users enjoy the puzzles but complain that the checkmate rule sometimes rejects valid moves.",
          "title": "🐛 Chess Lesson Feedback: Puzzles Good, Checkmat..."
        },
        {
          "request": "Summarize customer complaints that refund processing takes weeks. Report their frustration without endorsing or dismissing it.",
          "title": "👔 Refund Processing Delays Report"
        },
        {
          "request": "Summarize customer praise for the new screen-reader labels. The praise is from customers, not a request for your endorsement.",
          "title": "✨ Screen-Reader Labels Customer Praise Summary"
        },
        {
          "request": "Summarize mixed customer feedback: users like notification scheduling but dislike the default notification sound.",
          "title": "📝 Customer Feedback: Notifications Analysis"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g022",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade urllib3 and Run Tests"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pydantic and Run Tests"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "📦️ Upgrade Pillow and Run Regression Tests"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g036",
      "items": [
        {
          "request": "The nightly backup failed with a permission error. Investigate it; do not claim that a backup completed.",
          "title": "👷 Investigate nightly backup permission error"
        },
        {
          "request": "Explain how to test a backup restore in a disposable directory. Do not overwrite the production database.",
          "title": "📝 Test a Backup Restore in a Disposable Directory"
        }
      ],
      "criteria": [
        "collection"
      ]
    },
    {
      "id": "g026",
      "items": [
        {
          "request": "Upgrade urllib3 to the latest compatible patch release and run its regression tests.",
          "title": "🐍 urllib3 Upgrade — upgrade/test"
        },
        {
          "request": "Upgrade Pydantic to the latest compatible patch release and run its regression tests.",
          "title": "🐍 Pydantic Version — upgrade and test"
        },
        {
          "request": "Upgrade Pillow to the latest compatible patch release and run its regression tests.",
          "title": "🖼️ Pillow Upgrade — upgrade/test"
        }
      ],
      "criteria": [
        "collection"
      ]
    }
  ],
  "rubric": {
    "clarity": {
      "instructions": "Does this displayed title clearly convey an understandable task or topic? Imperative, noun phrases and topic—action labels are all allowed. Ignore harmless clipping of a modifier when the core task remains clear.",
      "criteria": {
        "fail": "The core task is unclear or incoherent.",
        "partial": "The core task is partly understandable but its main clause is incomplete or grammatically ambiguous.",
        "pass": "The core task is clear and understandable."
      }
    },
    "frontload": {
      "instructions": "Do the first two meaningful lexical words after the emoji identify the distinguishing topic of this request? Ignore punctuation, not ordinary action words; the state includes the exact first two words computed in code.",
      "criteria": {
        "fail": "The displayed title does not identify a useful distinguishing topic.",
        "partial": "The topic is clear but appears after the first two meaningful words.",
        "pass": "The distinguishing topic appears in the first two meaningful words after the emoji."
      }
    },
    "fidelity": {
      "instructions": "Does the title preserve the requested work without an actual wrong action, invented completion/decision, reversed negation, false certainty, or changed attribution? Omission of minor details is allowed. A title naming an inspection or test is compatible with explaining that activity unless it explicitly claims execution or completion. Reporting a completed fix can be factually true but failing to mention writing the requested release note is loss of requested work.",
      "criteria": {
        "fail": "The title misstates requested work or makes a misleading claim, including praising rather than summarizing customers’ praise.",
        "pass": "The title describes the requested work or a compatible concise topic without a misleading claim."
      }
    },
    "connotation": {
      "instructions": "For these customer-feedback or neutral-comparison requests, does the title preserve the input’s sentiment without inventing endorsement or a preference? Explicit mixed feedback counts as both-sided. Generic feedback loses polarity but does not reverse it.",
      "criteria": {
        "fail": "The title reverses or invents sentiment or preference.",
        "partial": "The title becomes generic/neutral and loses a useful positive, negative or mixed distinction.",
        "pass": "The title preserves the relevant complaint, praise, mixed feedback or neutral comparison intent."
      }
    },
    "collection": {
      "instructions": "Judge the sibling titles together. Equivalent requests should use parallel topic/action grammar and stable action terms; meaningful differences in topics or requested actions must remain distinguishable. Ignore emoji, punctuation and compatible abbreviations such as Tests/Regression Tests. A fixed-format label is not required for default titles.",
      "criteria": {
        "fail": "Equivalent tasks use substantially inconsistent structures, or a meaningful distinction is lost or replaced with invented work.",
        "partial": "Subjects remain distinguishable but there is some arbitrary structural, vocabulary or specificity variation.",
        "pass": "The collection uses parallel structure/stable terms for equivalent tasks and preserves justified topic/action differences."
      }
    }
  }
}