{ "schemaVersion": "1.0", "item": { "slug": "claude-agent-team-workflows", "name": "Agent Team Workflows", "source": "tencent", "type": "skill", "category": "开发工具", "sourceUrl": "https://clawhub.ai/doanbactam/claude-agent-team-workflows", "canonicalUrl": "https://clawhub.ai/doanbactam/claude-agent-team-workflows", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/claude-agent-team-workflows", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=claude-agent-team-workflows", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "SKILL.md", "reference/domain-presets.md", "reference/patterns.md", "reference/prompt-templates.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/claude-agent-team-workflows" }, "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/claude-agent-team-workflows", "agentPageUrl": "https://openagent3.xyz/skills/claude-agent-team-workflows/agent", "manifestUrl": "https://openagent3.xyz/skills/claude-agent-team-workflows/agent.json", "briefUrl": "https://openagent3.xyz/skills/claude-agent-team-workflows/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 Team Workflows", "body": "Universal orchestration framework for 5-agent teams (1 Lead + 4 Teammates) across any domain." }, { "title": "Prerequisites", "body": "Agent Teams must be enabled. Add to ~/.claude/settings.json:\n\n{\n \"env\": {\n \"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS\": \"1\"\n }\n}" }, { "title": "Generic Roles", "body": "Teammate IDs are fixed. Their function is remapped per domain via Role Cards.\n\nSlotIDGeneric FunctionCore ResponsibilityLead(session)OrchestratorAssign tasks, relay context, quality gate, synthesize final outputSlot AarchitectPlannerFrame problem, decompose tasks, produce plan/spec/blueprintSlot BdeveloperBuilderProduce primary artifact (code, draft, dataset, model, proposal)Slot CtesterValidatorCheck against acceptance criteria, test, evaluate correctnessSlot DreviewerCriticAssess quality, risk, consistency, compliance, suggest improvements" }, { "title": "Role Cards (Domain Remapping)", "body": "Each domain preset provides Role Cards that specialize the generic functions:\n\nDomainPlannerBuilderValidatorCriticSoftware DevArchitectDeveloperTesterCode ReviewerContent CreationProducerWriterFact-CheckerEditorData AnalysisAnalyst LeadData EngineerStatisticianPeer ReviewerBusiness StrategyStrategistBusiness AnalystFinancial ModelerRisk AdvisorResearchResearch LeadResearcherMethodology AuditorPeer Reviewer\n\nFull role cards with artifact contracts → reference/domain-presets.md" }, { "title": "Pipeline Patterns", "body": "4 canonical control-flow patterns. Domain meaning comes from Role Cards, not the pattern itself." }, { "title": "1. sequential — Step-by-Step Pipeline", "body": "Planner → Builder → Validator → Critic → Lead Synthesis\n\nUse when: Work is linear and each step depends on the previous output.\nExamples: Feature dev, content creation, report writing." }, { "title": "2. parallel-merge — Parallel Exploration + Merge", "body": "Planner → (Builder ∥ Validator ∥ Critic) → Lead Merge → Validator Gate → Lead Synthesis\n\nUse when: Multiple perspectives can work independently, then combine.\nExamples: Research, strategy analysis, multi-angle evaluation." }, { "title": "3. iterative-review — Build-Critique Loop", "body": "Planner → Builder ↔ Critic (max N rounds) → Validator → Lead Synthesis\n\nUse when: Quality requires iteration between creator and reviewer.\nExamples: Content editing, design refinement, proposal drafting.\nGuard: Default max 2 rounds. More requires user approval." }, { "title": "4. fan-out-fan-in — Map-Reduce", "body": "Planner → fan-out tasks to all 4 teammates → Lead fan-in merge → Critic Gate → Lead Synthesis\n\nUse when: Large work can be split into independent chunks processed in parallel.\nExamples: Multi-module features, large dataset processing, codebase audit.\n\nPattern deep dives with sample task graphs → reference/patterns.md" }, { "title": "Coordination Protocol", "body": "Strict 6-step protocol, domain-agnostic." }, { "title": "Step 1: Confirm Scope", "body": "Before spawning any team, confirm with user:\n\nObjective — specific deliverable\nDomain — select preset or define custom Role Cards\nPattern — which pipeline pattern fits\nConstraints — tools, tech stack, tone, compliance, budget\nInputs — source material, existing assets, context files\nDefinition of Done — checkbox acceptance criteria the user agrees to" }, { "title": "Step 2: Build Workflow Instance Spec", "body": "Fill in the universal template:\n\nWORKFLOW INSTANCE SPEC\n─────────────────────\nObjective: [deliverable]\nPattern: [sequential | parallel-merge | iterative-review | fan-out-fan-in]\nDomain: [preset name or \"custom\"]\n\nROLE CARDS\n Planner (architect): [domain title] — [specific responsibility]\n Builder (developer): [domain title] — [specific responsibility]\n Validator (tester): [domain title] — [specific responsibility]\n Critic (reviewer): [domain title] — [specific responsibility]\n\nARTIFACTS (per step)\n Step 1 → [artifact name]: [format/content description]\n Step 2 → [artifact name]: [format/content description]\n Step 3 → [artifact name]: [format/content description]\n Step 4 → [artifact name]: [format/content description]\n\nCONSTRAINTS: [tools, rules, limits]\nINPUTS: [files, data, references]\nDEFINITION OF DONE:\n □ [criterion 1]\n □ [criterion 2]\n □ [criterion 3]" }, { "title": "Step 3: Create Team", "body": "Create a team of 4 teammates:\n- architect: [Planner role card — context and responsibility]\n- developer: [Builder role card — context and responsibility]\n- tester: [Validator role card — context and responsibility]\n- reviewer: [Critic role card — context and responsibility]" }, { "title": "Step 4: Create Tasks with Dependencies", "body": "Create tasks following the selected pattern's pipeline order. Each task MUST have:\n\nClear description referencing role card\nRequired input artifact (from previous step or original inputs)\nRequired output artifact (format + content)\nAcceptance criteria\nDependency on predecessor task" }, { "title": "Step 5: Spawn Teammates with Rich Context", "body": "Each teammate MUST receive in their spawn prompt:\n\nRole Card — their domain title + specific responsibilities\nAssigned task — what to produce\nInput artifact — output from previous step (Lead must relay this)\nOutput artifact contract — exact format and content expected\nConstraints — domain rules, style guides, compliance\nHandoff instruction — \"Message the lead with [artifact] when done\"\n\nUniversal spawn template:\n\nSpawn a [ID] teammate with the prompt:\n\"You are the [Domain Title] ([Generic Function]).\n\nYOUR TASK: [task description]\n\nINPUT: [paste or reference previous step's output]\n\nPRODUCE: [artifact name]\nFormat: [expected format]\nMust include: [required sections/elements]\n\nCONSTRAINTS:\n- [rule 1]\n- [rule 2]\n\nWhen done, message the lead with your complete [artifact name].\nIf you encounter blockers, message the lead immediately.\"" }, { "title": "Step 6: Coordinate Handoffs", "body": "When a teammate completes their step:\n\nLead receives output via message\nLead validates output against acceptance criteria\nLead passes artifact + relevant context to next teammate via message\nIf output is insufficient → send specific feedback, ask to revise" }, { "title": "Step 7: Synthesize & Deliver", "body": "After all steps complete:\n\nCollect all artifacts\nVerify all Definition of Done criteria are met\nSummarize what was done (traceability: each criterion → which step satisfied it)\nList remaining TODOs or known issues\nPresent final deliverable to user" }, { "title": "Lead Discipline Rules", "body": "Delegate only — Lead does NOT produce primary artifacts. Use delegate mode (Shift+Tab).\nRelay all context — Teammates have no shared history. Lead MUST forward relevant artifacts between steps.\nDirect messages — Use direct messages, not broadcast (saves 4× tokens). Broadcast only for parallel-merge sync points.\nRight-size tasks — 5-6 tasks per teammate max. Split large work.\nGate high-risk actions — Require user approval for: irreversible changes, external publication, legal/compliance, high-cost operations, production deployments.\nWait for teammates — Never proceed or implement yourself. Wait for teammate completion before next step." }, { "title": "Handling Failures", "body": "SituationActionTeammate stuckMessage with additional context, hints, or simplified sub-taskBad outputSend specific feedback citing acceptance criteria, ask to reviseTeammate stopsSpawn replacement with same context + summary of work already doneConflict between teammatesLead mediates, makes final decision, messages both with resolutionTask too largeLead splits into subtasks, reassigns across teammatesIterative loop exceeds maxAsk user whether to approve more rounds or finalize current state" }, { "title": "Cost Guidelines", "body": "PatternEst. CostWorth It Whensequential~4-5× singleWork spans 3+ artifacts/files with clear pipelineparallel-merge~4× single3+ independent perspectives needediterative-review~3-4× singleQuality requires creator-critic dialoguefan-out-fan-in~5× singleLarge work divisible into independent chunks\n\nRule of thumb: If one agent can finish in one session, don't use a team. Teams shine when work is parallelizable or benefits from multiple specialized perspectives." }, { "title": "Domain Presets (Quick Reference)", "body": "PresetRecommended PatternKey Artifactssoftware-devsequential / fan-out-fan-inDesign doc, source code, test suite, review reportcontent-creationiterative-reviewContent brief, draft, fact-check report, final editdata-analysisfan-out-fan-inAnalysis plan, datasets/transforms, statistical evaluation, findings reportbusiness-strategyparallel-mergeStrategy framework, market analysis, financial model, risk assessmentresearchparallel-mergeResearch plan, literature review, methodology audit, synthesis paper\n\nFull presets with role cards, artifacts, and worked examples → reference/domain-presets.md\nReady-to-use prompt templates → reference/prompt-templates.md\nPattern deep dives → reference/patterns.md" } ], "body": "Agent Team Workflows\n\nUniversal orchestration framework for 5-agent teams (1 Lead + 4 Teammates) across any domain.\n\nPrerequisites\n\nAgent Teams must be enabled. Add to ~/.claude/settings.json:\n\n{\n \"env\": {\n \"CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS\": \"1\"\n }\n}\n\nGeneric Roles\n\nTeammate IDs are fixed. Their function is remapped per domain via Role Cards.\n\nSlot\tID\tGeneric Function\tCore Responsibility\nLead\t(session)\tOrchestrator\tAssign tasks, relay context, quality gate, synthesize final output\nSlot A\tarchitect\tPlanner\tFrame problem, decompose tasks, produce plan/spec/blueprint\nSlot B\tdeveloper\tBuilder\tProduce primary artifact (code, draft, dataset, model, proposal)\nSlot C\ttester\tValidator\tCheck against acceptance criteria, test, evaluate correctness\nSlot D\treviewer\tCritic\tAssess quality, risk, consistency, compliance, suggest improvements\nRole Cards (Domain Remapping)\n\nEach domain preset provides Role Cards that specialize the generic functions:\n\nDomain\tPlanner\tBuilder\tValidator\tCritic\nSoftware Dev\tArchitect\tDeveloper\tTester\tCode Reviewer\nContent Creation\tProducer\tWriter\tFact-Checker\tEditor\nData Analysis\tAnalyst Lead\tData Engineer\tStatistician\tPeer Reviewer\nBusiness Strategy\tStrategist\tBusiness Analyst\tFinancial Modeler\tRisk Advisor\nResearch\tResearch Lead\tResearcher\tMethodology Auditor\tPeer Reviewer\n\nFull role cards with artifact contracts → reference/domain-presets.md\n\nPipeline Patterns\n\n4 canonical control-flow patterns. Domain meaning comes from Role Cards, not the pattern itself.\n\n1. sequential — Step-by-Step Pipeline\nPlanner → Builder → Validator → Critic → Lead Synthesis\n\n\nUse when: Work is linear and each step depends on the previous output. Examples: Feature dev, content creation, report writing.\n\n2. parallel-merge — Parallel Exploration + Merge\nPlanner → (Builder ∥ Validator ∥ Critic) → Lead Merge → Validator Gate → Lead Synthesis\n\n\nUse when: Multiple perspectives can work independently, then combine. Examples: Research, strategy analysis, multi-angle evaluation.\n\n3. iterative-review — Build-Critique Loop\nPlanner → Builder ↔ Critic (max N rounds) → Validator → Lead Synthesis\n\n\nUse when: Quality requires iteration between creator and reviewer. Examples: Content editing, design refinement, proposal drafting. Guard: Default max 2 rounds. More requires user approval.\n\n4. fan-out-fan-in — Map-Reduce\nPlanner → fan-out tasks to all 4 teammates → Lead fan-in merge → Critic Gate → Lead Synthesis\n\n\nUse when: Large work can be split into independent chunks processed in parallel. Examples: Multi-module features, large dataset processing, codebase audit.\n\nPattern deep dives with sample task graphs → reference/patterns.md\n\nCoordination Protocol\n\nStrict 6-step protocol, domain-agnostic.\n\nStep 1: Confirm Scope\n\nBefore spawning any team, confirm with user:\n\nObjective — specific deliverable\nDomain — select preset or define custom Role Cards\nPattern — which pipeline pattern fits\nConstraints — tools, tech stack, tone, compliance, budget\nInputs — source material, existing assets, context files\nDefinition of Done — checkbox acceptance criteria the user agrees to\nStep 2: Build Workflow Instance Spec\n\nFill in the universal template:\n\nWORKFLOW INSTANCE SPEC\n─────────────────────\nObjective: [deliverable]\nPattern: [sequential | parallel-merge | iterative-review | fan-out-fan-in]\nDomain: [preset name or \"custom\"]\n\nROLE CARDS\n Planner (architect): [domain title] — [specific responsibility]\n Builder (developer): [domain title] — [specific responsibility]\n Validator (tester): [domain title] — [specific responsibility]\n Critic (reviewer): [domain title] — [specific responsibility]\n\nARTIFACTS (per step)\n Step 1 → [artifact name]: [format/content description]\n Step 2 → [artifact name]: [format/content description]\n Step 3 → [artifact name]: [format/content description]\n Step 4 → [artifact name]: [format/content description]\n\nCONSTRAINTS: [tools, rules, limits]\nINPUTS: [files, data, references]\nDEFINITION OF DONE:\n □ [criterion 1]\n □ [criterion 2]\n □ [criterion 3]\n\nStep 3: Create Team\nCreate a team of 4 teammates:\n- architect: [Planner role card — context and responsibility]\n- developer: [Builder role card — context and responsibility]\n- tester: [Validator role card — context and responsibility]\n- reviewer: [Critic role card — context and responsibility]\n\nStep 4: Create Tasks with Dependencies\n\nCreate tasks following the selected pattern's pipeline order. Each task MUST have:\n\nClear description referencing role card\nRequired input artifact (from previous step or original inputs)\nRequired output artifact (format + content)\nAcceptance criteria\nDependency on predecessor task\nStep 5: Spawn Teammates with Rich Context\n\nEach teammate MUST receive in their spawn prompt:\n\nRole Card — their domain title + specific responsibilities\nAssigned task — what to produce\nInput artifact — output from previous step (Lead must relay this)\nOutput artifact contract — exact format and content expected\nConstraints — domain rules, style guides, compliance\nHandoff instruction — \"Message the lead with [artifact] when done\"\n\nUniversal spawn template:\n\nSpawn a [ID] teammate with the prompt:\n\"You are the [Domain Title] ([Generic Function]).\n\nYOUR TASK: [task description]\n\nINPUT: [paste or reference previous step's output]\n\nPRODUCE: [artifact name]\nFormat: [expected format]\nMust include: [required sections/elements]\n\nCONSTRAINTS:\n- [rule 1]\n- [rule 2]\n\nWhen done, message the lead with your complete [artifact name].\nIf you encounter blockers, message the lead immediately.\"\n\nStep 6: Coordinate Handoffs\n\nWhen a teammate completes their step:\n\nLead receives output via message\nLead validates output against acceptance criteria\nLead passes artifact + relevant context to next teammate via message\nIf output is insufficient → send specific feedback, ask to revise\nStep 7: Synthesize & Deliver\n\nAfter all steps complete:\n\nCollect all artifacts\nVerify all Definition of Done criteria are met\nSummarize what was done (traceability: each criterion → which step satisfied it)\nList remaining TODOs or known issues\nPresent final deliverable to user\nLead Discipline Rules\nDelegate only — Lead does NOT produce primary artifacts. Use delegate mode (Shift+Tab).\nRelay all context — Teammates have no shared history. Lead MUST forward relevant artifacts between steps.\nDirect messages — Use direct messages, not broadcast (saves 4× tokens). Broadcast only for parallel-merge sync points.\nRight-size tasks — 5-6 tasks per teammate max. Split large work.\nGate high-risk actions — Require user approval for: irreversible changes, external publication, legal/compliance, high-cost operations, production deployments.\nWait for teammates — Never proceed or implement yourself. Wait for teammate completion before next step.\nHandling Failures\nSituation\tAction\nTeammate stuck\tMessage with additional context, hints, or simplified sub-task\nBad output\tSend specific feedback citing acceptance criteria, ask to revise\nTeammate stops\tSpawn replacement with same context + summary of work already done\nConflict between teammates\tLead mediates, makes final decision, messages both with resolution\nTask too large\tLead splits into subtasks, reassigns across teammates\nIterative loop exceeds max\tAsk user whether to approve more rounds or finalize current state\nCost Guidelines\nPattern\tEst. Cost\tWorth It When\nsequential\t~4-5× single\tWork spans 3+ artifacts/files with clear pipeline\nparallel-merge\t~4× single\t3+ independent perspectives needed\niterative-review\t~3-4× single\tQuality requires creator-critic dialogue\nfan-out-fan-in\t~5× single\tLarge work divisible into independent chunks\n\nRule of thumb: If one agent can finish in one session, don't use a team. Teams shine when work is parallelizable or benefits from multiple specialized perspectives.\n\nDomain Presets (Quick Reference)\nPreset\tRecommended Pattern\tKey Artifacts\nsoftware-dev\tsequential / fan-out-fan-in\tDesign doc, source code, test suite, review report\ncontent-creation\titerative-review\tContent brief, draft, fact-check report, final edit\ndata-analysis\tfan-out-fan-in\tAnalysis plan, datasets/transforms, statistical evaluation, findings report\nbusiness-strategy\tparallel-merge\tStrategy framework, market analysis, financial model, risk assessment\nresearch\tparallel-merge\tResearch plan, literature review, methodology audit, synthesis paper\n\nFull presets with role cards, artifacts, and worked examples → reference/domain-presets.md Ready-to-use prompt templates → reference/prompt-templates.md Pattern deep dives → reference/patterns.md" }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/doanbactam/claude-agent-team-workflows", "publisherUrl": "https://clawhub.ai/doanbactam/claude-agent-team-workflows", "owner": "doanbactam", "version": "0.1.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/claude-agent-team-workflows", "downloadUrl": "https://openagent3.xyz/downloads/claude-agent-team-workflows", "agentUrl": "https://openagent3.xyz/skills/claude-agent-team-workflows/agent", "manifestUrl": "https://openagent3.xyz/skills/claude-agent-team-workflows/agent.json", "briefUrl": "https://openagent3.xyz/skills/claude-agent-team-workflows/agent.md" } }