{ "schemaVersion": "1.0", "item": { "slug": "agent-swarm-workflow", "name": "Agent Swarm Workflow", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/clawdnw/agent-swarm-workflow", "canonicalUrl": "https://clawhub.ai/clawdnw/agent-swarm-workflow", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/agent-swarm-workflow", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=agent-swarm-workflow", "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/agent-swarm-workflow" }, "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-swarm-workflow", "agentPageUrl": "https://openagent3.xyz/skills/agent-swarm-workflow/agent", "manifestUrl": "https://openagent3.xyz/skills/agent-swarm-workflow/agent.json", "briefUrl": "https://openagent3.xyz/skills/agent-swarm-workflow/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 Swarm Workflow — Parallel Implementation", "body": "Core Insight: Every agent is fungible and a generalist. They all use the same base model and read the same AGENTS.md. Simply telling one it's a \"frontend agent\" doesn't make it better at frontend.\nThe swarm is distributed, robust, and self-organizing through Agent Mail and Beads." }, { "title": "Prerequisites", "body": "Before starting a swarm:\n\nComprehensive plan created (see planning-workflow skill)\nPolished beads ready (see beads-workflow skill)\nAGENTS.md configured with all tool blurbs\nAgent Mail server running (am or ~/projects/mcp_agent_mail/scripts/run_server_with_token.sh)\nNTM available for session management" }, { "title": "The Swarm Architecture", "body": "┌─────────────────────────────────────────────────────────────┐\n│ BEADS │\n│ (Task graph with dependencies, priorities, status) │\n└─────────────────────────────────────────────────────────────┘\n │\n ┌─────────┴─────────┐\n ▼ ▼\n┌─────────────────────────────┐ ┌─────────────────────────┐\n│ BV │ │ AGENT MAIL │\n│ (What to work on) │ │ (Coordination layer) │\n└─────────────────────────────┘ └─────────────────────────┘\n │ │\n └──────────────┬─────────────┘\n ▼\n┌─────────────────────────────────────────────────────────────┐\n│ NTM + AGENTS │\n│ ┌─────┐ ┌─────┐ ┌─────┐ ┌─────┐ ┌─────┐ ┌─────┐ │\n│ │ CC │ │ CC │ │ Cod │ │ Gmi │ │ CC │ │ Cod │ │\n│ └─────┘ └─────┘ └─────┘ └─────┘ └─────┘ └─────┘ │\n└─────────────────────────────────────────────────────────────┘" }, { "title": "Using NTM (Named Tmux Manager)", "body": "# Spawn a swarm with multiple agents\nntm spawn myproject --cc=3 --cod=2 --gmi=1\n\n# Send initial prompt to all Claude Code agents\nntm send myproject --cc \"$(cat initial_prompt.txt)\"\n\n# Or send to all agents\nntm send myproject --all \"$(cat initial_prompt.txt)\"" }, { "title": "Manual Setup", "body": "Create tmux sessions/panes for each agent in your project folder." }, { "title": "THE EXACT PROMPT — Initial Agent Marching Orders", "body": "Give each agent this EXACT prompt to start:\n\nFirst read ALL of the AGENTS dot md file and README dot md file super carefully and understand ALL of both! Then use your code investigation agent mode to fully understand the code, and technical architecture and purpose of the project. Then register with MCP Agent Mail and introduce yourself to the other agents.\n\nBe sure to check your agent mail and to promptly respond if needed to any messages; then proceed meticulously with your next assigned beads, working on the tasks systematically and meticulously and tracking your progress via beads and agent mail messages.\n\nDon't get stuck in \"communication purgatory\" where nothing is getting done; be proactive about starting tasks that need to be done, but inform your fellow agents via messages when you do so and mark beads appropriately.\n\nWhen you're not sure what to do next, use the bv tool mentioned in AGENTS dot md to prioritize the best beads to work on next; pick the next one that you can usefully work on and get started. Make sure to acknowledge all communication requests from other agents and that you are aware of all active agents and their names. Use ultrathink." }, { "title": "THE EXACT PROMPT — Move to Next Bead", "body": "Once agents complete a bead, use this prompt to keep them moving:\n\nReread AGENTS dot md so it's still fresh in your mind. Use ultrathink. Use bv with the robot flags (see AGENTS dot md for info on this) to find the most impactful bead(s) to work on next and then start on it. Remember to mark the beads appropriately and communicate with your fellow agents. Pick the next bead you can actually do usefully now and start coding on it immediately; communicate what you're working on to your fellow agents and mark beads appropriately as you work. And respond to any agent mail messages you've received." }, { "title": "THE EXACT PROMPT — Self-Review After Bead Completion", "body": "Have agents review their work before moving on:\n\nGreat, now I want you to carefully read over all of the new code you just wrote and other existing code you just modified with \"fresh eyes\" looking super carefully for any obvious bugs, errors, problems, issues, confusion, etc. Carefully fix anything you uncover. Use ultrathink.\n\nKeep running this until they stop finding bugs." }, { "title": "THE EXACT PROMPT — Post-Compaction", "body": "When an agent does a compaction, immediately follow up with:\n\nReread AGENTS dot md so it's still fresh in your mind. Use ultrathink.\n\nThis re-establishes the critical context about tools and workflows." }, { "title": "THE EXACT PROMPT — Cross-Agent Review", "body": "Periodically have agents review each other's work:\n\nOk can you now turn your attention to reviewing the code written by your fellow agents and checking for any issues, bugs, errors, problems, inefficiencies, security problems, reliability issues, etc. and carefully diagnose their underlying root causes using first-principle analysis and then fix or revise them if necessary? Don't restrict yourself to the latest commits, cast a wider net and go super deep! Use ultrathink." }, { "title": "THE EXACT PROMPT — Random Code Exploration", "body": "For deep quality checks:\n\nI want you to sort of randomly explore the code files in this project, choosing code files to deeply investigate and understand and trace their functionality and execution flows through the related code files which they import or which they are imported by.\n\nOnce you understand the purpose of the code in the larger context of the workflows, I want you to do a super careful, methodical, and critical check with \"fresh eyes\" to find any obvious bugs, problems, errors, issues, silly mistakes, etc. and then systematically and meticulously and intelligently correct them.\n\nBe sure to comply with ALL rules in AGENTS dot md and ensure that any code you write or revise conforms to the best practice guides referenced in the AGENTS dot md file. Use ultrathink." }, { "title": "THE EXACT PROMPT — Commit Changes", "body": "Have agents commit logically grouped changes:\n\nNow, based on your knowledge of the project, commit all changed files now in a series of logically connected groupings with super detailed commit messages for each and then push. Take your time to do it right. Don't edit the code at all. Don't commit obviously ephemeral files. Use ultrathink." }, { "title": "THE EXACT PROMPT — Add Test Coverage", "body": "Do we have full unit test coverage without using mocks/fake stuff? What about complete e2e integration test scripts with great, detailed logging? If not, then create a comprehensive and granular set of beads for all this with tasks, subtasks, and dependency structure overlaid with detailed comments." }, { "title": "THE EXACT PROMPT — UI/UX Scrutiny", "body": "Great, now I want you to super carefully scrutinize every aspect of the application workflow and implementation and look for things that just seem sub-optimal or even wrong/mistaken to you, things that could very obviously be improved from a user-friendliness and intuitiveness standpoint, places where our UI/UX could be improved and polished to be slicker, more visually appealing, and more premium feeling and just ultra high quality, like Stripe-level apps." }, { "title": "THE EXACT PROMPT — Deep UI/UX Enhancement", "body": "I still think there are strong opportunities to enhance the UI/UX look and feel and to make everything work better and be more intuitive, user-friendly, visually appealing, polished, slick, and world class in terms of following UI/UX best practices like those used by Stripe, don't you agree? And I want you to carefully consider desktop UI/UX and mobile UI/UX separately while doing this and hyper-optimize for both separately to play to the specifics of each modality. I'm looking for true world-class visual appeal, polish, slickness, etc. that makes people gasp at how stunning and perfect it is in every way. Use ultrathink." }, { "title": "How Agents Coordinate", "body": "Each agent:\n\nRegisters with Agent Mail at session start\nReserves files before editing (file_reservation_paths)\nAnnounces work via messages with bead ID in thread_id\nChecks inbox between tasks\nReleases reservations when done" }, { "title": "File Reservation Pattern", "body": "# Before starting work on a bead\nfile_reservation_paths(\n project_key=\"/path/to/project\",\n agent_name=\"GreenCastle\",\n paths=[\"src/auth/**/*.ts\"],\n ttl_seconds=3600,\n exclusive=True,\n reason=\"bd-123\"\n)\n\n# After completing work\nrelease_file_reservations(project_key, agent_name)" }, { "title": "Communication Pattern", "body": "# Announce starting a bead\nsend_message(\n project_key, agent_name,\n to=[\"BlueLake\", \"RedMountain\"], # Other active agents\n subject=\"[bd-123] Starting auth module\",\n body_md=\"I'm taking bd-123. Reserved src/auth/**.\",\n thread_id=\"bd-123\"\n)\n\n# Update on completion\nsend_message(\n project_key, agent_name,\n to=[\"BlueLake\", \"RedMountain\"],\n subject=\"[bd-123] Completed\",\n body_md=\"Auth module done and tested. Released file reservations.\",\n thread_id=\"bd-123\"\n)" }, { "title": "Key Commands", "body": "# THE MEGA-COMMAND: Start here\nbv --robot-triage\n\n# Just get the single top pick\nbv --robot-next\n\n# Get parallel execution tracks (for multi-agent)\nbv --robot-plan\n\n# Check for cycles (MUST FIX if found)\nbv --robot-insights | jq '.Cycles'\n\n# Find bottlenecks\nbv --robot-insights | jq '.bottlenecks'\n\nCRITICAL: Never run bare bv — it launches interactive TUI that blocks the session." }, { "title": "Run Until Clean", "body": "Keep running these prompts in rounds until they consistently come back with no changes:\n\nSelf-review — Agent reviews own code\nCross-review — Agent reviews other agents' code\nRandom exploration — Deep dive into random code paths" }, { "title": "The Steady-State Signal", "body": "When all three types of reviews return clean (no bugs found, no changes made), the code is likely solid." }, { "title": "Initial Marching Orders", "body": "First read ALL of the AGENTS dot md file and README dot md file super carefully and understand ALL of both! Then use your code investigation agent mode to fully understand the code, and technical architecture and purpose of the project. Then register with MCP Agent Mail and introduce yourself to the other agents.\n\nBe sure to check your agent mail and to promptly respond if needed to any messages; then proceed meticulously with your next assigned beads, working on the tasks systematically and meticulously and tracking your progress via beads and agent mail messages.\n\nDon't get stuck in \"communication purgatory\" where nothing is getting done; be proactive about starting tasks that need to be done, but inform your fellow agents via messages when you do so and mark beads appropriately.\n\nWhen you're not sure what to do next, use the bv tool mentioned in AGENTS dot md to prioritize the best beads to work on next; pick the next one that you can usefully work on and get started. Make sure to acknowledge all communication requests from other agents and that you are aware of all active agents and their names. Use ultrathink." }, { "title": "Move to Next Bead", "body": "Reread AGENTS dot md so it's still fresh in your mind. Use ultrathink. Use bv with the robot flags (see AGENTS dot md for info on this) to find the most impactful bead(s) to work on next and then start on it. Remember to mark the beads appropriately and communicate with your fellow agents. Pick the next bead you can actually do usefully now and start coding on it immediately; communicate what you're working on to your fellow agents and mark beads appropriately as you work. And respond to any agent mail messages you've received." }, { "title": "Self-Review", "body": "Great, now I want you to carefully read over all of the new code you just wrote and other existing code you just modified with \"fresh eyes\" looking super carefully for any obvious bugs, errors, problems, issues, confusion, etc. Carefully fix anything you uncover. Use ultrathink." }, { "title": "Cross-Review", "body": "Ok can you now turn your attention to reviewing the code written by your fellow agents and checking for any issues, bugs, errors, problems, inefficiencies, security problems, reliability issues, etc. and carefully diagnose their underlying root causes using first-principle analysis and then fix or revise them if necessary? Don't restrict yourself to the latest commits, cast a wider net and go super deep! Use ultrathink." }, { "title": "Random Exploration", "body": "I want you to sort of randomly explore the code files in this project, choosing code files to deeply investigate and understand and trace their functionality and execution flows through the related code files which they import or which they are imported by.\n\nOnce you understand the purpose of the code in the larger context of the workflows, I want you to do a super careful, methodical, and critical check with \"fresh eyes\" to find any obvious bugs, problems, errors, issues, silly mistakes, etc. and then systematically and meticulously and intelligently correct them.\n\nBe sure to comply with ALL rules in AGENTS dot md and ensure that any code you write or revises conforms to the best practice guides referenced in the AGENTS dot md file. Use ultrathink." }, { "title": "Post-Compaction", "body": "Reread AGENTS dot md so it's still fresh in your mind. Use ultrathink." }, { "title": "Commit Changes", "body": "Now, based on your knowledge of the project, commit all changed files now in a series of logically connected groupings with super detailed commit messages for each and then push. Take your time to do it right. Don't edit the code at all. Don't commit obviously ephemeral files. Use ultrathink." }, { "title": "Test Coverage", "body": "Do we have full unit test coverage without using mocks/fake stuff? What about complete e2e integration test scripts with great, detailed logging? If not, then create a comprehensive and granular set of beads for all this with tasks, subtasks, and dependency structure overlaid with detailed comments." }, { "title": "UI/UX Scrutiny", "body": "Great, now I want you to super carefully scrutinize every aspect of the application workflow and implementation and look for things that just seem sub-optimal or even wrong/mistaken to you, things that could very obviously be improved from a user-friendliness and intuitiveness standpoint, places where our UI/UX could be improved and polished to be slicker, more visually appealing, and more premium feeling and just ultra high quality, like Stripe-level apps." }, { "title": "Deep UI/UX Enhancement", "body": "I still think there are strong opportunities to enhance the UI/UX look and feel and to make everything work better and be more intuitive, user-friendly, visually appealing, polished, slick, and world class in terms of following UI/UX best practices like those used by Stripe, don't you agree? And I want you to carefully consider desktop UI/UX and mobile UI/UX separately while doing this and hyper-optimize for both separately to play to the specifics of each modality. I'm looking for true world-class visual appeal, polish, slickness, etc. that makes people gasp at how stunning and perfect it is in every way. Use ultrathink." }, { "title": "The Flywheel in Action", "body": "PLAN ──► BEADS ──► SWARM ──► REVIEW ──► COMMIT\n │ │ │ │\n │ │ └────┬────┘\n │ │ │\n │ └───── REPEAT ──┘\n │ │\n │ v2 PLAN ◄────────────┘\n │ │\n └──────────────┘\n\nEach cycle improves:\n\nCASS remembers solutions\nCM distills patterns\nUBS catches more issues\nBV shows graph health" }, { "title": "FAQ", "body": "Q: How do agents know what to work on?\nA: They use bv --robot-triage or bv --robot-next to find the highest-impact ready bead.\n\nQ: How do they avoid conflicts?\nA: File reservations in Agent Mail. Exclusive reservations block others; the pre-commit guard enforces it.\n\nQ: What if an agent crashes or forgets?\nA: Every agent is fungible. Start a new session, read AGENTS.md, check bead status, continue.\n\nQ: How many agents should I run?\nA: Depends on project complexity. Start with 3-6. More agents = faster but more coordination overhead.\n\nQ: What model mix works best?\nA: Mix recommended. Try 3 Claude Code (Opus), 2 Codex (GPT 5.2), 1 Gemini. They have different strengths.\n\nQ: Do agents need individual areas of expertise?\nA: No, every agent is fungible and a generalist. Simply telling one it's a frontend agent doesn't make it better at frontend.\n\nQ: Is there traceability between git commit and bead?\nA: Yes, bv automatically does this analysis and links beads to relevant git commits by analyzing the stream of data and making logical deductions." } ], "body": "Agent Swarm Workflow — Parallel Implementation\n\nCore Insight: Every agent is fungible and a generalist. They all use the same base model and read the same AGENTS.md. Simply telling one it's a \"frontend agent\" doesn't make it better at frontend.\n\nThe swarm is distributed, robust, and self-organizing through Agent Mail and Beads.\n\nPrerequisites\n\nBefore starting a swarm:\n\nComprehensive plan created (see planning-workflow skill)\nPolished beads ready (see beads-workflow skill)\nAGENTS.md configured with all tool blurbs\nAgent Mail server running (am or ~/projects/mcp_agent_mail/scripts/run_server_with_token.sh)\nNTM available for session management\nThe Swarm Architecture\n┌─────────────────────────────────────────────────────────────┐\n│ BEADS │\n│ (Task graph with dependencies, priorities, status) │\n└─────────────────────────────────────────────────────────────┘\n │\n ┌─────────┴─────────┐\n ▼ ▼\n┌─────────────────────────────┐ ┌─────────────────────────┐\n│ BV │ │ AGENT MAIL │\n│ (What to work on) │ │ (Coordination layer) │\n└─────────────────────────────┘ └─────────────────────────┘\n │ │\n └──────────────┬─────────────┘\n ▼\n┌─────────────────────────────────────────────────────────────┐\n│ NTM + AGENTS │\n│ ┌─────┐ ┌─────┐ ┌─────┐ ┌─────┐ ┌─────┐ ┌─────┐ │\n│ │ CC │ │ CC │ │ Cod │ │ Gmi │ │ CC │ │ Cod │ │\n│ └─────┘ └─────┘ └─────┘ └─────┘ └─────┘ └─────┘ │\n└─────────────────────────────────────────────────────────────┘\n\nStarting the Swarm\nUsing NTM (Named Tmux Manager)\n# Spawn a swarm with multiple agents\nntm spawn myproject --cc=3 --cod=2 --gmi=1\n\n# Send initial prompt to all Claude Code agents\nntm send myproject --cc \"$(cat initial_prompt.txt)\"\n\n# Or send to all agents\nntm send myproject --all \"$(cat initial_prompt.txt)\"\n\nManual Setup\n\nCreate tmux sessions/panes for each agent in your project folder.\n\nTHE EXACT PROMPT — Initial Agent Marching Orders\n\nGive each agent this EXACT prompt to start:\n\nFirst read ALL of the AGENTS dot md file and README dot md file super carefully and understand ALL of both! Then use your code investigation agent mode to fully understand the code, and technical architecture and purpose of the project. Then register with MCP Agent Mail and introduce yourself to the other agents.\n\nBe sure to check your agent mail and to promptly respond if needed to any messages; then proceed meticulously with your next assigned beads, working on the tasks systematically and meticulously and tracking your progress via beads and agent mail messages.\n\nDon't get stuck in \"communication purgatory\" where nothing is getting done; be proactive about starting tasks that need to be done, but inform your fellow agents via messages when you do so and mark beads appropriately.\n\nWhen you're not sure what to do next, use the bv tool mentioned in AGENTS dot md to prioritize the best beads to work on next; pick the next one that you can usefully work on and get started. Make sure to acknowledge all communication requests from other agents and that you are aware of all active agents and their names. Use ultrathink.\n\nThe Implementation Loop\nTHE EXACT PROMPT — Move to Next Bead\n\nOnce agents complete a bead, use this prompt to keep them moving:\n\nReread AGENTS dot md so it's still fresh in your mind. Use ultrathink. Use bv with the robot flags (see AGENTS dot md for info on this) to find the most impactful bead(s) to work on next and then start on it. Remember to mark the beads appropriately and communicate with your fellow agents. Pick the next bead you can actually do usefully now and start coding on it immediately; communicate what you're working on to your fellow agents and mark beads appropriately as you work. And respond to any agent mail messages you've received.\n\nTHE EXACT PROMPT — Self-Review After Bead Completion\n\nHave agents review their work before moving on:\n\nGreat, now I want you to carefully read over all of the new code you just wrote and other existing code you just modified with \"fresh eyes\" looking super carefully for any obvious bugs, errors, problems, issues, confusion, etc. Carefully fix anything you uncover. Use ultrathink.\n\n\nKeep running this until they stop finding bugs.\n\nHandling Context Compaction\nTHE EXACT PROMPT — Post-Compaction\n\nWhen an agent does a compaction, immediately follow up with:\n\nReread AGENTS dot md so it's still fresh in your mind. Use ultrathink.\n\n\nThis re-establishes the critical context about tools and workflows.\n\nQuality Review Prompts\nTHE EXACT PROMPT — Cross-Agent Review\n\nPeriodically have agents review each other's work:\n\nOk can you now turn your attention to reviewing the code written by your fellow agents and checking for any issues, bugs, errors, problems, inefficiencies, security problems, reliability issues, etc. and carefully diagnose their underlying root causes using first-principle analysis and then fix or revise them if necessary? Don't restrict yourself to the latest commits, cast a wider net and go super deep! Use ultrathink.\n\nTHE EXACT PROMPT — Random Code Exploration\n\nFor deep quality checks:\n\nI want you to sort of randomly explore the code files in this project, choosing code files to deeply investigate and understand and trace their functionality and execution flows through the related code files which they import or which they are imported by.\n\nOnce you understand the purpose of the code in the larger context of the workflows, I want you to do a super careful, methodical, and critical check with \"fresh eyes\" to find any obvious bugs, problems, errors, issues, silly mistakes, etc. and then systematically and meticulously and intelligently correct them.\n\nBe sure to comply with ALL rules in AGENTS dot md and ensure that any code you write or revise conforms to the best practice guides referenced in the AGENTS dot md file. Use ultrathink.\n\nCommitting Work\nTHE EXACT PROMPT — Commit Changes\n\nHave agents commit logically grouped changes:\n\nNow, based on your knowledge of the project, commit all changed files now in a series of logically connected groupings with super detailed commit messages for each and then push. Take your time to do it right. Don't edit the code at all. Don't commit obviously ephemeral files. Use ultrathink.\n\nPost-Bead Completion Prompts\nTHE EXACT PROMPT — Add Test Coverage\nDo we have full unit test coverage without using mocks/fake stuff? What about complete e2e integration test scripts with great, detailed logging? If not, then create a comprehensive and granular set of beads for all this with tasks, subtasks, and dependency structure overlaid with detailed comments.\n\nTHE EXACT PROMPT — UI/UX Scrutiny\nGreat, now I want you to super carefully scrutinize every aspect of the application workflow and implementation and look for things that just seem sub-optimal or even wrong/mistaken to you, things that could very obviously be improved from a user-friendliness and intuitiveness standpoint, places where our UI/UX could be improved and polished to be slicker, more visually appealing, and more premium feeling and just ultra high quality, like Stripe-level apps.\n\nTHE EXACT PROMPT — Deep UI/UX Enhancement\nI still think there are strong opportunities to enhance the UI/UX look and feel and to make everything work better and be more intuitive, user-friendly, visually appealing, polished, slick, and world class in terms of following UI/UX best practices like those used by Stripe, don't you agree? And I want you to carefully consider desktop UI/UX and mobile UI/UX separately while doing this and hyper-optimize for both separately to play to the specifics of each modality. I'm looking for true world-class visual appeal, polish, slickness, etc. that makes people gasp at how stunning and perfect it is in every way. Use ultrathink.\n\nAgent Mail Integration\nHow Agents Coordinate\n\nEach agent:\n\nRegisters with Agent Mail at session start\nReserves files before editing (file_reservation_paths)\nAnnounces work via messages with bead ID in thread_id\nChecks inbox between tasks\nReleases reservations when done\nFile Reservation Pattern\n# Before starting work on a bead\nfile_reservation_paths(\n project_key=\"/path/to/project\",\n agent_name=\"GreenCastle\",\n paths=[\"src/auth/**/*.ts\"],\n ttl_seconds=3600,\n exclusive=True,\n reason=\"bd-123\"\n)\n\n# After completing work\nrelease_file_reservations(project_key, agent_name)\n\nCommunication Pattern\n# Announce starting a bead\nsend_message(\n project_key, agent_name,\n to=[\"BlueLake\", \"RedMountain\"], # Other active agents\n subject=\"[bd-123] Starting auth module\",\n body_md=\"I'm taking bd-123. Reserved src/auth/**.\",\n thread_id=\"bd-123\"\n)\n\n# Update on completion\nsend_message(\n project_key, agent_name,\n to=[\"BlueLake\", \"RedMountain\"],\n subject=\"[bd-123] Completed\",\n body_md=\"Auth module done and tested. Released file reservations.\",\n thread_id=\"bd-123\"\n)\n\nUsing BV for Task Selection\nKey Commands\n# THE MEGA-COMMAND: Start here\nbv --robot-triage\n\n# Just get the single top pick\nbv --robot-next\n\n# Get parallel execution tracks (for multi-agent)\nbv --robot-plan\n\n# Check for cycles (MUST FIX if found)\nbv --robot-insights | jq '.Cycles'\n\n# Find bottlenecks\nbv --robot-insights | jq '.bottlenecks'\n\n\nCRITICAL: Never run bare bv — it launches interactive TUI that blocks the session.\n\nQuality Loops\nRun Until Clean\n\nKeep running these prompts in rounds until they consistently come back with no changes:\n\nSelf-review — Agent reviews own code\nCross-review — Agent reviews other agents' code\nRandom exploration — Deep dive into random code paths\nThe Steady-State Signal\n\nWhen all three types of reviews return clean (no bugs found, no changes made), the code is likely solid.\n\nComplete Prompt Reference\nInitial Marching Orders\nFirst read ALL of the AGENTS dot md file and README dot md file super carefully and understand ALL of both! Then use your code investigation agent mode to fully understand the code, and technical architecture and purpose of the project. Then register with MCP Agent Mail and introduce yourself to the other agents.\n\nBe sure to check your agent mail and to promptly respond if needed to any messages; then proceed meticulously with your next assigned beads, working on the tasks systematically and meticulously and tracking your progress via beads and agent mail messages.\n\nDon't get stuck in \"communication purgatory\" where nothing is getting done; be proactive about starting tasks that need to be done, but inform your fellow agents via messages when you do so and mark beads appropriately.\n\nWhen you're not sure what to do next, use the bv tool mentioned in AGENTS dot md to prioritize the best beads to work on next; pick the next one that you can usefully work on and get started. Make sure to acknowledge all communication requests from other agents and that you are aware of all active agents and their names. Use ultrathink.\n\nMove to Next Bead\nReread AGENTS dot md so it's still fresh in your mind. Use ultrathink. Use bv with the robot flags (see AGENTS dot md for info on this) to find the most impactful bead(s) to work on next and then start on it. Remember to mark the beads appropriately and communicate with your fellow agents. Pick the next bead you can actually do usefully now and start coding on it immediately; communicate what you're working on to your fellow agents and mark beads appropriately as you work. And respond to any agent mail messages you've received.\n\nSelf-Review\nGreat, now I want you to carefully read over all of the new code you just wrote and other existing code you just modified with \"fresh eyes\" looking super carefully for any obvious bugs, errors, problems, issues, confusion, etc. Carefully fix anything you uncover. Use ultrathink.\n\nCross-Review\nOk can you now turn your attention to reviewing the code written by your fellow agents and checking for any issues, bugs, errors, problems, inefficiencies, security problems, reliability issues, etc. and carefully diagnose their underlying root causes using first-principle analysis and then fix or revise them if necessary? Don't restrict yourself to the latest commits, cast a wider net and go super deep! Use ultrathink.\n\nRandom Exploration\nI want you to sort of randomly explore the code files in this project, choosing code files to deeply investigate and understand and trace their functionality and execution flows through the related code files which they import or which they are imported by.\n\nOnce you understand the purpose of the code in the larger context of the workflows, I want you to do a super careful, methodical, and critical check with \"fresh eyes\" to find any obvious bugs, problems, errors, issues, silly mistakes, etc. and then systematically and meticulously and intelligently correct them.\n\nBe sure to comply with ALL rules in AGENTS dot md and ensure that any code you write or revises conforms to the best practice guides referenced in the AGENTS dot md file. Use ultrathink.\n\nPost-Compaction\nReread AGENTS dot md so it's still fresh in your mind. Use ultrathink.\n\nCommit Changes\nNow, based on your knowledge of the project, commit all changed files now in a series of logically connected groupings with super detailed commit messages for each and then push. Take your time to do it right. Don't edit the code at all. Don't commit obviously ephemeral files. Use ultrathink.\n\nTest Coverage\nDo we have full unit test coverage without using mocks/fake stuff? What about complete e2e integration test scripts with great, detailed logging? If not, then create a comprehensive and granular set of beads for all this with tasks, subtasks, and dependency structure overlaid with detailed comments.\n\nUI/UX Scrutiny\nGreat, now I want you to super carefully scrutinize every aspect of the application workflow and implementation and look for things that just seem sub-optimal or even wrong/mistaken to you, things that could very obviously be improved from a user-friendliness and intuitiveness standpoint, places where our UI/UX could be improved and polished to be slicker, more visually appealing, and more premium feeling and just ultra high quality, like Stripe-level apps.\n\nDeep UI/UX Enhancement\nI still think there are strong opportunities to enhance the UI/UX look and feel and to make everything work better and be more intuitive, user-friendly, visually appealing, polished, slick, and world class in terms of following UI/UX best practices like those used by Stripe, don't you agree? And I want you to carefully consider desktop UI/UX and mobile UI/UX separately while doing this and hyper-optimize for both separately to play to the specifics of each modality. I'm looking for true world-class visual appeal, polish, slickness, etc. that makes people gasp at how stunning and perfect it is in every way. Use ultrathink.\n\nThe Flywheel in Action\nPLAN ──► BEADS ──► SWARM ──► REVIEW ──► COMMIT\n │ │ │ │\n │ │ └────┬────┘\n │ │ │\n │ └───── REPEAT ──┘\n │ │\n │ v2 PLAN ◄────────────┘\n │ │\n └──────────────┘\n\n\nEach cycle improves:\n\nCASS remembers solutions\nCM distills patterns\nUBS catches more issues\nBV shows graph health\nFAQ\n\nQ: How do agents know what to work on? A: They use bv --robot-triage or bv --robot-next to find the highest-impact ready bead.\n\nQ: How do they avoid conflicts? A: File reservations in Agent Mail. Exclusive reservations block others; the pre-commit guard enforces it.\n\nQ: What if an agent crashes or forgets? A: Every agent is fungible. Start a new session, read AGENTS.md, check bead status, continue.\n\nQ: How many agents should I run? A: Depends on project complexity. Start with 3-6. More agents = faster but more coordination overhead.\n\nQ: What model mix works best? A: Mix recommended. Try 3 Claude Code (Opus), 2 Codex (GPT 5.2), 1 Gemini. They have different strengths.\n\nQ: Do agents need individual areas of expertise? A: No, every agent is fungible and a generalist. Simply telling one it's a frontend agent doesn't make it better at frontend.\n\nQ: Is there traceability between git commit and bead? A: Yes, bv automatically does this analysis and links beads to relevant git commits by analyzing the stream of data and making logical deductions." }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/clawdnw/agent-swarm-workflow", "publisherUrl": "https://clawhub.ai/clawdnw/agent-swarm-workflow", "owner": "clawdnw", "version": "1.0.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/agent-swarm-workflow", "downloadUrl": "https://openagent3.xyz/downloads/agent-swarm-workflow", "agentUrl": "https://openagent3.xyz/skills/agent-swarm-workflow/agent", "manifestUrl": "https://openagent3.xyz/skills/agent-swarm-workflow/agent.json", "briefUrl": "https://openagent3.xyz/skills/agent-swarm-workflow/agent.md" } }