{
  "schemaVersion": "1.0",
  "item": {
    "slug": "agent-hive",
    "name": "Agent Hive",
    "source": "tencent",
    "type": "skill",
    "category": "效率提升",
    "sourceUrl": "https://clawhub.ai/nissan/agent-hive",
    "canonicalUrl": "https://clawhub.ai/nissan/agent-hive",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/agent-hive",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=agent-hive",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "SKILL.md",
      "references/governance.md",
      "references/heartbeat-snippet.md",
      "scripts/budget_audit.py",
      "scripts/create_agent.sh"
    ],
    "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/agent-hive"
    },
    "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/agent-hive",
    "agentPageUrl": "https://openagent3.xyz/skills/agent-hive/agent",
    "manifestUrl": "https://openagent3.xyz/skills/agent-hive/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/agent-hive/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": "Agent Hive",
        "body": "Create and manage multi-agent teams with shared memory, budget governance, and configurable spawn permissions."
      },
      {
        "title": "Architecture",
        "body": "Main Workspace (~/.openclaw/workspace/)\n├── MEMORY.md, TOOLS.md, USER.md, IDENTITY.md   ← shared brain (real files)\n├── agents/<name>/                                ← per-agent outbox, inbox, budget\n├── agents/governance/                            ← governance rules + audit log\n├── memory/, projects/, scripts/, skills/         ← shared resources\n└── content/                                      ← shared content\n\nAgent Workspace (~/.openclaw/workspace-<id>/)\n├── SOUL.md              ← unique personality (local file)\n├── AGENTS.md            ← unique instructions (local file)\n├── HEARTBEAT.md         ← agent-specific heartbeat (local file)\n├── .openclaw/           ← agent-specific config dir\n└── everything else      ← SYMLINKS to main workspace"
      },
      {
        "title": "Step 1: Create workspace with symlinks",
        "body": "AGENT_ID=\"<id>\"\nMAIN=\"$HOME/.openclaw/workspace\"\nWS=\"$HOME/.openclaw/workspace-$AGENT_ID\"\n\nmkdir -p \"$WS/.openclaw\"\n\n# Symlink shared brain\nfor f in .learnings IDENTITY.md MEMORY.md ROADMAP.md TOOLS.md USER.md; do\n  ln -sf \"../workspace/$f\" \"$WS/$f\"\ndone\nfor d in agents content memory projects scripts skills; do\n  ln -sf \"../workspace/$d\" \"$WS/$d\"\ndone"
      },
      {
        "title": "Step 2: Create unique files",
        "body": "Create $WS/SOUL.md — the agent's personality. This is theirs alone.\n\nCreate $WS/AGENTS.md — agent-specific instructions. Must include:\n\nSession startup checklist (read SOUL.md, check BUDGET.json, check INBOX.md)\nCommunication rules (INBOX/OUTBOX pattern)\nWhat the agent does and doesn't do\nBudget rules section (see references/budget-rules.md)\n\nCreate $WS/HEARTBEAT.md — minimal heartbeat config."
      },
      {
        "title": "Step 3: Create agent directory in shared workspace",
        "body": "mkdir -p \"$MAIN/agents/$AGENT_ID\"\ntouch \"$MAIN/agents/$AGENT_ID/INBOX.md\"\ntouch \"$MAIN/agents/$AGENT_ID/OUTBOX.md\""
      },
      {
        "title": "Step 4: Create budget file",
        "body": "Copy scripts/create_budget.sh output or create manually:\n\n{\n  \"daily_limit_output_tokens\": 50000,\n  \"today\": \"YYYY-MM-DD\",\n  \"used_output_tokens\": 0,\n  \"spawns\": [],\n  \"status\": \"active\",\n  \"warnings\": [],\n  \"consecutive_overbudget_days\": 0\n}\n\nSave to $MAIN/agents/$AGENT_ID/BUDGET.json."
      },
      {
        "title": "Step 5: Register in openclaw.json",
        "body": "Add to agents.list[]:\n\n{\n  \"id\": \"<id>\",\n  \"name\": \"<Name>\",\n  \"workspace\": \"/absolute/path/to/workspace-<id>\",\n  \"identity\": { \"name\": \"<Name>\", \"emoji\": \"<emoji>\" },\n  \"model\": {\n    \"primary\": \"anthropic/claude-sonnet-4-6\",\n    \"fallbacks\": [\"<fallback-model-1>\", \"<fallback-model-2>\"]\n  },\n  \"subagents\": { \"allowAgents\": [\"<peer1>\", \"<peer2>\"] }\n}\n\nUpdate existing agents' allowAgents to include the new agent."
      },
      {
        "title": "Step 6: Validate and restart",
        "body": "python3 -c \"import json; json.load(open('$HOME/.openclaw/openclaw.json')); print('JSON valid')\"\nopenclaw doctor  # check for schema errors\nlaunchctl stop ai.openclaw.gateway && sleep 2 && launchctl start ai.openclaw.gateway\n\nWait 15 seconds before testing."
      },
      {
        "title": "Hub-and-spoke (conservative)",
        "body": "Only the orchestrator (main) can spawn agents. Agents route requests through OUTBOX.\n\n\"subagents\": { \"allowAgents\": [] }  // for all non-main agents"
      },
      {
        "title": "Full mesh with budget (recommended)",
        "body": "All agents can spawn peers. Budget governance prevents abuse.\n\n// Each agent can spawn all peers (not themselves)\n\"subagents\": { \"allowAgents\": [\"<all other agent ids>\"] }"
      },
      {
        "title": "Earned mesh (progressive trust)",
        "body": "Start hub-and-spoke. Promote to mesh after demonstrating budget responsibility.\nDemote back to hub-and-spoke after repeated overspend."
      },
      {
        "title": "Budget Governance",
        "body": "See references/governance.md for the full framework.\n\nRun the audit script on heartbeat:\n\npython3 scripts/budget_audit.py\n\nThresholds: Green (<80%), Yellow (80-100%), Red (>100%), Emergency demotion (>200%).\n3 consecutive overbudget days → automatic demotion (mesh privileges revoked).\n\nAdd step 4.5 to your HEARTBEAT.md — see references/heartbeat-snippet.md."
      },
      {
        "title": "Committee Pattern",
        "body": "For project work, compose agent committees:\n\nIdentify which agents are needed (e.g., content + marketing + design review)\nSpawn them in parallel with clear, scoped tasks\nEach writes to their OUTBOX.md\nOrchestrator collects results and synthesizes\nBudget tracked per-agent across the project"
      },
      {
        "title": "Verification Checklist",
        "body": "After adding an agent, verify:\n\nworkspace-<id>/ has SOUL.md + AGENTS.md as local files\n All other files are symlinks to main workspace\n agents/<id>/BUDGET.json exists with correct schema\n Agent appears in openclaw doctor output\n Gateway restarts without errors\n Agent can be spawned: sessions_spawn(agentId=\"<id>\", task=\"Say hello\")\n Budget audit includes the new agent: python3 scripts/budget_audit.py"
      }
    ],
    "body": "Agent Hive\n\nCreate and manage multi-agent teams with shared memory, budget governance, and configurable spawn permissions.\n\nArchitecture\nMain Workspace (~/.openclaw/workspace/)\n├── MEMORY.md, TOOLS.md, USER.md, IDENTITY.md   ← shared brain (real files)\n├── agents/<name>/                                ← per-agent outbox, inbox, budget\n├── agents/governance/                            ← governance rules + audit log\n├── memory/, projects/, scripts/, skills/         ← shared resources\n└── content/                                      ← shared content\n\nAgent Workspace (~/.openclaw/workspace-<id>/)\n├── SOUL.md              ← unique personality (local file)\n├── AGENTS.md            ← unique instructions (local file)\n├── HEARTBEAT.md         ← agent-specific heartbeat (local file)\n├── .openclaw/           ← agent-specific config dir\n└── everything else      ← SYMLINKS to main workspace\n\nAdding a New Agent\nStep 1: Create workspace with symlinks\nAGENT_ID=\"<id>\"\nMAIN=\"$HOME/.openclaw/workspace\"\nWS=\"$HOME/.openclaw/workspace-$AGENT_ID\"\n\nmkdir -p \"$WS/.openclaw\"\n\n# Symlink shared brain\nfor f in .learnings IDENTITY.md MEMORY.md ROADMAP.md TOOLS.md USER.md; do\n  ln -sf \"../workspace/$f\" \"$WS/$f\"\ndone\nfor d in agents content memory projects scripts skills; do\n  ln -sf \"../workspace/$d\" \"$WS/$d\"\ndone\n\nStep 2: Create unique files\n\nCreate $WS/SOUL.md — the agent's personality. This is theirs alone.\n\nCreate $WS/AGENTS.md — agent-specific instructions. Must include:\n\nSession startup checklist (read SOUL.md, check BUDGET.json, check INBOX.md)\nCommunication rules (INBOX/OUTBOX pattern)\nWhat the agent does and doesn't do\nBudget rules section (see references/budget-rules.md)\n\nCreate $WS/HEARTBEAT.md — minimal heartbeat config.\n\nStep 3: Create agent directory in shared workspace\nmkdir -p \"$MAIN/agents/$AGENT_ID\"\ntouch \"$MAIN/agents/$AGENT_ID/INBOX.md\"\ntouch \"$MAIN/agents/$AGENT_ID/OUTBOX.md\"\n\nStep 4: Create budget file\n\nCopy scripts/create_budget.sh output or create manually:\n\n{\n  \"daily_limit_output_tokens\": 50000,\n  \"today\": \"YYYY-MM-DD\",\n  \"used_output_tokens\": 0,\n  \"spawns\": [],\n  \"status\": \"active\",\n  \"warnings\": [],\n  \"consecutive_overbudget_days\": 0\n}\n\n\nSave to $MAIN/agents/$AGENT_ID/BUDGET.json.\n\nStep 5: Register in openclaw.json\n\nAdd to agents.list[]:\n\n{\n  \"id\": \"<id>\",\n  \"name\": \"<Name>\",\n  \"workspace\": \"/absolute/path/to/workspace-<id>\",\n  \"identity\": { \"name\": \"<Name>\", \"emoji\": \"<emoji>\" },\n  \"model\": {\n    \"primary\": \"anthropic/claude-sonnet-4-6\",\n    \"fallbacks\": [\"<fallback-model-1>\", \"<fallback-model-2>\"]\n  },\n  \"subagents\": { \"allowAgents\": [\"<peer1>\", \"<peer2>\"] }\n}\n\n\nUpdate existing agents' allowAgents to include the new agent.\n\nStep 6: Validate and restart\npython3 -c \"import json; json.load(open('$HOME/.openclaw/openclaw.json')); print('JSON valid')\"\nopenclaw doctor  # check for schema errors\nlaunchctl stop ai.openclaw.gateway && sleep 2 && launchctl start ai.openclaw.gateway\n\n\nWait 15 seconds before testing.\n\nSpawn Permission Models\nHub-and-spoke (conservative)\n\nOnly the orchestrator (main) can spawn agents. Agents route requests through OUTBOX.\n\n\"subagents\": { \"allowAgents\": [] }  // for all non-main agents\n\nFull mesh with budget (recommended)\n\nAll agents can spawn peers. Budget governance prevents abuse.\n\n// Each agent can spawn all peers (not themselves)\n\"subagents\": { \"allowAgents\": [\"<all other agent ids>\"] }\n\nEarned mesh (progressive trust)\n\nStart hub-and-spoke. Promote to mesh after demonstrating budget responsibility. Demote back to hub-and-spoke after repeated overspend.\n\nBudget Governance\n\nSee references/governance.md for the full framework.\n\nRun the audit script on heartbeat:\n\npython3 scripts/budget_audit.py\n\n\nThresholds: Green (<80%), Yellow (80-100%), Red (>100%), Emergency demotion (>200%). 3 consecutive overbudget days → automatic demotion (mesh privileges revoked).\n\nAdd step 4.5 to your HEARTBEAT.md — see references/heartbeat-snippet.md.\n\nCommittee Pattern\n\nFor project work, compose agent committees:\n\nIdentify which agents are needed (e.g., content + marketing + design review)\nSpawn them in parallel with clear, scoped tasks\nEach writes to their OUTBOX.md\nOrchestrator collects results and synthesizes\nBudget tracked per-agent across the project\nVerification Checklist\n\nAfter adding an agent, verify:\n\n workspace-<id>/ has SOUL.md + AGENTS.md as local files\n All other files are symlinks to main workspace\n agents/<id>/BUDGET.json exists with correct schema\n Agent appears in openclaw doctor output\n Gateway restarts without errors\n Agent can be spawned: sessions_spawn(agentId=\"<id>\", task=\"Say hello\")\n Budget audit includes the new agent: python3 scripts/budget_audit.py"
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/nissan/agent-hive",
    "publisherUrl": "https://clawhub.ai/nissan/agent-hive",
    "owner": "nissan",
    "version": "1.0.1",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/agent-hive",
    "downloadUrl": "https://openagent3.xyz/downloads/agent-hive",
    "agentUrl": "https://openagent3.xyz/skills/agent-hive/agent",
    "manifestUrl": "https://openagent3.xyz/skills/agent-hive/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/agent-hive/agent.md"
  }
}