{
  "schemaVersion": "1.0",
  "item": {
    "slug": "cassandra",
    "name": "Cassandra",
    "source": "tencent",
    "type": "skill",
    "category": "AI 智能",
    "sourceUrl": "https://clawhub.ai/ivangdavila/cassandra",
    "canonicalUrl": "https://clawhub.ai/ivangdavila/cassandra",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/cassandra",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=cassandra",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "SKILL.md"
    ],
    "primaryDoc": "SKILL.md",
    "quickSetup": [
      "Download the package from Yavira.",
      "Extract the archive and review SKILL.md first.",
      "Import or place the package into your OpenClaw setup."
    ],
    "agentAssist": {
      "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
      "steps": [
        "Download the package from Yavira.",
        "Extract it into a folder your agent can access.",
        "Paste one of the prompts below and point your agent at the extracted folder."
      ],
      "prompts": [
        {
          "label": "New install",
          "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete."
        },
        {
          "label": "Upgrade existing",
          "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run."
        }
      ]
    },
    "sourceHealth": {
      "source": "tencent",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-04-23T16:43:11.935Z",
      "expiresAt": "2026-04-30T16:43:11.935Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=4claw-imageboard",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=4claw-imageboard",
        "contentDisposition": "attachment; filename=\"4claw-imageboard-1.0.1.zip\"",
        "redirectLocation": null,
        "bodySnippet": null
      },
      "scope": "source",
      "summary": "Source download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this source.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/cassandra"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    },
    "downloadPageUrl": "https://openagent3.xyz/downloads/cassandra",
    "agentPageUrl": "https://openagent3.xyz/skills/cassandra/agent",
    "manifestUrl": "https://openagent3.xyz/skills/cassandra/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/cassandra/agent.md"
  },
  "agentAssist": {
    "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
    "steps": [
      "Download the package from Yavira.",
      "Extract it into a folder your agent can access.",
      "Paste one of the prompts below and point your agent at the extracted folder."
    ],
    "prompts": [
      {
        "label": "New install",
        "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete."
      },
      {
        "label": "Upgrade existing",
        "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run."
      }
    ]
  },
  "documentation": {
    "source": "clawhub",
    "primaryDoc": "SKILL.md",
    "sections": [
      {
        "title": "Data Modeling Mistakes",
        "body": "Design tables around queries, not entities—denormalization is mandatory, not optional\nOne table per query pattern—Cassandra has no JOINs; duplicate data across tables\nPartition key determines data distribution—all rows with same partition key on same node\nWide partitions kill performance—keep under 100MB; add time bucket to partition key if growing"
      },
      {
        "title": "Primary Key Traps",
        "body": "PRIMARY KEY (a, b, c): a is partition key, b and c are clustering columns\nPRIMARY KEY ((a, b), c): (a, b) together is partition key—compound partition key\nClustering columns define sort order within partition—query must respect this order\nCan't query by clustering column without partition key—unlike SQL indexes"
      },
      {
        "title": "Query Restrictions",
        "body": "WHERE must include full partition key—partial partition key fails unless ALLOW FILTERING\nALLOW FILTERING scans all nodes—never use in production; redesign table instead\nRange queries only on last clustering column used—WHERE a = ? AND b > ? works, WHERE a = ? AND c > ? doesn't\nIN on partition key hits multiple nodes—expensive; prefer single partition queries"
      },
      {
        "title": "Consistency Levels",
        "body": "QUORUM for most operations—majority of replicas; balances consistency and availability\nLOCAL_QUORUM for multi-datacenter—avoids cross-DC latency\nONE for pure availability—may read stale data; fine for caches, bad for critical reads\nWrite + read consistency must overlap for strong consistency—QUORUM + QUORUM safe"
      },
      {
        "title": "Tombstones (Silent Performance Killer)",
        "body": "DELETE creates a tombstone, not actual deletion—tombstones persist until compaction\nMass deletes destroy read performance—thousands of tombstones scanned per query\nTTL also creates tombstones—don't use short TTLs with high write volume\nCheck with nodetool cfstats -H table—Tombstone columns show problem"
      },
      {
        "title": "Batch Misuse",
        "body": "UNLOGGED BATCH is not faster—use only for atomic writes to same partition\nLOGGED BATCH for multi-partition atomicity—adds coordination overhead\nDon't batch unrelated writes—hurts coordinator; send individual async writes\nBatch size limit ~50KB—larger batches fail or timeout"
      },
      {
        "title": "Anti-Patterns",
        "body": "Secondary indexes on high-cardinality columns—scatter-gather query, slow\nSecondary indexes on frequently updated columns—creates tombstones\nSELECT *—always list columns; schema changes break queries\nUUID as partition key without time component—random distribution, hot spots during bulk loads"
      },
      {
        "title": "Lightweight Transactions",
        "body": "IF NOT EXISTS / IF column = ?—uses Paxos, 4x slower than normal write\nSerial consistency for LWTs—SERIAL or LOCAL_SERIAL\nDon't use for counters or high-frequency updates—contention kills throughput\nReturns [applied] boolean—must check if operation succeeded"
      },
      {
        "title": "Collections and Counters",
        "body": "Sets/Lists/Maps stored with row—can't exceed 64KB, no pagination\nList prepend is anti-pattern—creates tombstones; use append or Set\nCounters require dedicated table—can't mix with regular columns\nCounter increment is not idempotent—retry may double-count"
      },
      {
        "title": "Compaction Strategies",
        "body": "SizeTieredCompactionStrategy (default)—good for write-heavy, uses more disk space\nLeveledCompactionStrategy—better read latency, higher write amplification\nTimeWindowCompactionStrategy—for time-series with TTL; reduces tombstone overhead\nWrong strategy for workload = degraded performance over time"
      },
      {
        "title": "Operations",
        "body": "nodetool repair regularly—inconsistencies accumulate without repair\nnodetool status shows cluster health—UN (Up Normal) is good, DN is down\nSchema changes propagate eventually—wait for nodetool describecluster to show agreement\nRolling restarts: one node at a time, wait for UN status before next"
      }
    ],
    "body": "Data Modeling Mistakes\nDesign tables around queries, not entities—denormalization is mandatory, not optional\nOne table per query pattern—Cassandra has no JOINs; duplicate data across tables\nPartition key determines data distribution—all rows with same partition key on same node\nWide partitions kill performance—keep under 100MB; add time bucket to partition key if growing\nPrimary Key Traps\nPRIMARY KEY (a, b, c): a is partition key, b and c are clustering columns\nPRIMARY KEY ((a, b), c): (a, b) together is partition key—compound partition key\nClustering columns define sort order within partition—query must respect this order\nCan't query by clustering column without partition key—unlike SQL indexes\nQuery Restrictions\nWHERE must include full partition key—partial partition key fails unless ALLOW FILTERING\nALLOW FILTERING scans all nodes—never use in production; redesign table instead\nRange queries only on last clustering column used—WHERE a = ? AND b > ? works, WHERE a = ? AND c > ? doesn't\nIN on partition key hits multiple nodes—expensive; prefer single partition queries\nConsistency Levels\nQUORUM for most operations—majority of replicas; balances consistency and availability\nLOCAL_QUORUM for multi-datacenter—avoids cross-DC latency\nONE for pure availability—may read stale data; fine for caches, bad for critical reads\nWrite + read consistency must overlap for strong consistency—QUORUM + QUORUM safe\nTombstones (Silent Performance Killer)\nDELETE creates a tombstone, not actual deletion—tombstones persist until compaction\nMass deletes destroy read performance—thousands of tombstones scanned per query\nTTL also creates tombstones—don't use short TTLs with high write volume\nCheck with nodetool cfstats -H table—Tombstone columns show problem\nBatch Misuse\nUNLOGGED BATCH is not faster—use only for atomic writes to same partition\nLOGGED BATCH for multi-partition atomicity—adds coordination overhead\nDon't batch unrelated writes—hurts coordinator; send individual async writes\nBatch size limit ~50KB—larger batches fail or timeout\nAnti-Patterns\nSecondary indexes on high-cardinality columns—scatter-gather query, slow\nSecondary indexes on frequently updated columns—creates tombstones\nSELECT *—always list columns; schema changes break queries\nUUID as partition key without time component—random distribution, hot spots during bulk loads\nLightweight Transactions\nIF NOT EXISTS / IF column = ?—uses Paxos, 4x slower than normal write\nSerial consistency for LWTs—SERIAL or LOCAL_SERIAL\nDon't use for counters or high-frequency updates—contention kills throughput\nReturns [applied] boolean—must check if operation succeeded\nCollections and Counters\nSets/Lists/Maps stored with row—can't exceed 64KB, no pagination\nList prepend is anti-pattern—creates tombstones; use append or Set\nCounters require dedicated table—can't mix with regular columns\nCounter increment is not idempotent—retry may double-count\nCompaction Strategies\nSizeTieredCompactionStrategy (default)—good for write-heavy, uses more disk space\nLeveledCompactionStrategy—better read latency, higher write amplification\nTimeWindowCompactionStrategy—for time-series with TTL; reduces tombstone overhead\nWrong strategy for workload = degraded performance over time\nOperations\nnodetool repair regularly—inconsistencies accumulate without repair\nnodetool status shows cluster health—UN (Up Normal) is good, DN is down\nSchema changes propagate eventually—wait for nodetool describecluster to show agreement\nRolling restarts: one node at a time, wait for UN status before next"
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/ivangdavila/cassandra",
    "publisherUrl": "https://clawhub.ai/ivangdavila/cassandra",
    "owner": "ivangdavila",
    "version": "1.0.0",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/cassandra",
    "downloadUrl": "https://openagent3.xyz/downloads/cassandra",
    "agentUrl": "https://openagent3.xyz/skills/cassandra/agent",
    "manifestUrl": "https://openagent3.xyz/skills/cassandra/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/cassandra/agent.md"
  }
}