[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EM6Z0WDW5WN2SFSTXY9VFD":3,"skill-same:01M7EM6Z0WDW5WN2SFSTXY9VFD":45},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":29,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7EM6Z0WDW5WN2SFSTXY9VFD","ams-agent-management",null,"TanNhatCMS\u002Fjarvis","TanNhatCMS","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F113796420?v=4","01M753MV17KG5CD3152G3WQEQY","KietNT","01M7EKEDYVYQ6PWVJJMZ911VMX",0,"MIT","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Fams-agent-management\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: agent-management\ndescription: Manage fleets of local AI agents: lifecycle, monitoring, task assignment, and resource control. Use when running multiple CLI agents via an orchestrator like AI Maestro.\ncategory: ai-maestro\n---\n\n# Agent Management\n\n## Overview\n\nRunning one AI agent is a conversation. Running many is operations: lifecycle (start\u002Fstop\u002Frestart), health monitoring, task assignment, and resource limits.\n\nThe AI Maestro pattern: local CLI agents run in managed sessions (e.g., tmux), supervised by an orchestrator that assigns work and watches health.\n\nGood management means: agents are visible, tasks are tracked, failures are noticed, and resources stay bounded.\n\n## When to use\n\n- Running multiple AI agents simultaneously\n- Agents dying silently or hanging unnoticed\n- Assigning and tracking work across agents\n- Long-running agent fleets needing supervision\n- Setting up local agent infrastructure\n\n## Core concepts\n\n- **Lifecycle management.**\n  Start, stop, restart, hibernate, wake. Agents are processes with states — manage them explicitly, not by hoping.\n- **Session isolation.**\n  Each agent in its own session (tmux\u002Fscreen): independent I\u002FO, survivable disconnects, per-agent logs. Never share a session between agents.\n- **Health monitoring.**\n  Heartbeat or output-activity checks: is the agent alive, working, or stuck? Stuck detection (no output for X minutes) triggers alerts or restarts.\n- **Task assignment.**\n  Work queue model: tasks with clear specs assigned to agents; status tracked (queued\u002Frunning\u002Fdone\u002Ffailed). No ambiguous 'go help with stuff.'\n- **Resource bounds.**\n  CPU\u002Fmemory limits, max concurrent agents, timeout per task. Unbounded agents consume the machine and your budget.\n- **Logging per agent.**\n  Separate logs per agent\u002Fsession, timestamped. Debugging a fleet without per-agent logs is archaeology.\n- **Naming and roles.**\n  Descriptive agent names tied to roles (researcher-1, builder-docs). You manage what you can identify.\n- **Graceful shutdown.**\n  Stop signals that let agents finish current work and save state. Kill -9 is for emergencies, not routine stops.\n\n## Practical workflow\n\n1. **Define agent roles.**\n   What types do you need? Researcher, builder, reviewer, monitor? One role per agent type; name accordingly.\n2. **Set up sessions.**\n   One managed session per agent with logging. Standardize the launch: same environment, same working directory conventions.\n3. **Configure monitoring.**\n   Health checks (activity\u002Fheartbeat), stuck detection thresholds, alert channel. Test alerts by simulating a stuck agent.\n4. **Create the work queue.**\n   Task format: goal, context, success criteria, timeout. Assignment: pull model (agents take next task) beats push for reliability.\n5. **Set resource limits.**\n   Max agents, per-agent timeouts, system resource caps. Document what happens at the limits.\n6. **Run and observe.**\n   Start small (2-3 agents), watch closely for a day. Tune timeouts, prompts, and queue discipline from real behavior.\n7. **Handle failures.**\n   Failed task -> logged, requeued or escalated per policy. Dead agent -> alert + restart. Document the runbooks.\n8. **Review and scale.**\n   Weekly: throughput, failure rate, stuck incidents. Scale agent count only when the management layer proves reliable.\n\n## Common pitfalls\n\n- **Fire-and-forget agents.**\n  Launching agents with no monitoring. Silent death is the default outcome of unmanaged agents.\n- **Shared sessions.**\n  Two agents in one terminal session: interleaved I\u002FO, unkillable individually, undebuggable. One agent, one session.\n- **No task specs.**\n  'Work on the docs' as a task. Vague tasks produce vague results — spec the goal, context, and done criteria.\n- **Unbounded runs.**\n  No timeouts: one stuck agent burns budget for 12 hours. Every task gets a timeout; every agent gets supervision.\n- **No per-agent logs.**\n  One shared log for five agents. When something breaks, you can't tell who did what.\n- **Manual everything.**\n  Starting\u002Fstopping\u002Fassigning all by hand. The orchestrator should handle lifecycle; humans handle exceptions.\n- **Scaling too fast.**\n  20 agents before managing 3 reliably. Management discipline first, fleet size second.\n- **Ignoring costs.**\n  Agents burn tokens and compute. Track spend per agent\u002Ftask; set budgets and alerts.\n","# Agent Management\n\n## Overview\n\nRunning one AI agent is a conversation. Running many is operations: lifecycle (start\u002Fstop\u002Frestart), health monitoring, task assignment, and resource limits.\n\nThe AI Maestro pattern: local CLI agents run in managed sessions (e.g., tmux), supervised by an orchestrator that assigns work and watches health.\n\nGood management means: agents are visible, tasks are tracked, failures are noticed, and resources stay bounded.\n\n## When to use\n\n- Running multiple AI agents simultaneously\n- Agents dying silently or hanging unnoticed\n- Assigning and tracking work across agents\n- Long-running agent fleets needing supervision\n- Setting up local agent infrastructure\n\n## Core concepts\n\n- **Lifecycle management.**\n  Start, stop, restart, hibernate, wake. Agents are processes with states — manage them explicitly, not by hoping.\n- **Session isolation.**\n  Each agent in its own session (tmux\u002Fscreen): independent I\u002FO, survivable disconnects, per-agent logs. Never share a session between agents.\n- **Health monitoring.**\n  Heartbeat or output-activity checks: is the agent alive, working, or stuck? Stuck detection (no output for X minutes) triggers alerts or restarts.\n- **Task assignment.**\n  Work queue model: tasks with clear specs assigned to agents; status tracked (queued\u002Frunning\u002Fdone\u002Ffailed). No ambiguous 'go help with stuff.'\n- **Resource bounds.**\n  CPU\u002Fmemory limits, max concurrent agents, timeout per task. Unbounded agents consume the machine and your budget.\n- **Logging per agent.**\n  Separate logs per agent\u002Fsession, timestamped. Debugging a fleet without per-agent logs is archaeology.\n- **Naming and roles.**\n  Descriptive agent names tied to roles (researcher-1, builder-docs). You manage what you can identify.\n- **Graceful shutdown.**\n  Stop signals that let agents finish current work and save state. Kill -9 is for emergencies, not routine stops.\n\n## Practical workflow\n\n1. **Define agent roles.**\n   What types do you need? Researcher, builder, reviewer, monitor? One role per agent type; name accordingly.\n2. **Set up sessions.**\n   One managed session per agent with logging. Standardize the launch: same environment, same working directory conventions.\n3. **Configure monitoring.**\n   Health checks (activity\u002Fheartbeat), stuck detection thresholds, alert channel. Test alerts by simulating a stuck agent.\n4. **Create the work queue.**\n   Task format: goal, context, success criteria, timeout. Assignment: pull model (agents take next task) beats push for reliability.\n5. **Set resource limits.**\n   Max agents, per-agent timeouts, system resource caps. Document what happens at the limits.\n6. **Run and observe.**\n   Start small (2-3 agents), watch closely for a day. Tune timeouts, prompts, and queue discipline from real behavior.\n7. **Handle failures.**\n   Failed task -> logged, requeued or escalated per policy. Dead agent -> alert + restart. Document the runbooks.\n8. **Review and scale.**\n   Weekly: throughput, failure rate, stuck incidents. Scale agent count only when the management layer proves reliable.\n\n## Common pitfalls\n\n- **Fire-and-forget agents.**\n  Launching agents with no monitoring. Silent death is the default outcome of unmanaged agents.\n- **Shared sessions.**\n  Two agents in one terminal session: interleaved I\u002FO, unkillable individually, undebuggable. One agent, one session.\n- **No task specs.**\n  'Work on the docs' as a task. Vague tasks produce vague results — spec the goal, context, and done criteria.\n- **Unbounded runs.**\n  No timeouts: one stuck agent burns budget for 12 hours. Every task gets a timeout; every agent gets supervision.\n- **No per-agent logs.**\n  One shared log for five agents. When something breaks, you can't tell who did what.\n- **Manual everything.**\n  Starting\u002Fstopping\u002Fassigning all by hand. The orchestrator should handle lifecycle; humans handle exceptions.\n- **Scaling too fast.**\n  20 agents before managing 3 reliably. Management discipline first, fleet size second.\n- **Ignoring costs.**\n  Agents burn tokens and compute. Track spend per agent\u002Ftask; set budgets and alerts.\n",{},"main","af388a77c2fe5e0b4fd879cca4300a0ffa9a6579","skills\u002Fams-agent-management",1,"https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis","https:\u002F\u002Fgithub.com\u002FTanNhatCMS","https:\u002F\u002Fmrkiet.name.vn","2026-11-07 14:23:08",{"source":7,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add TanNhatCMS\u002Fjarvis --skill ams-agent-management",{"tool":37,"command":38},"bunx","bunx skills add TanNhatCMS\u002Fjarvis --skill ams-agent-management",{"tool":40,"command":41},"pnpm","pnpm dlx skills add TanNhatCMS\u002Fjarvis --skill ams-agent-management","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Fams-agent-management\u002FSKILL.md",true,[],[46,53,61,68,75,82],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":18},"01M7EWYJR0J8DARY7E59Z6Y9MV","web-scrape","Turn web pages into LLM-ready markdown via the Firecrawl API (renders JS, anti-bot handling). Use when browser_open\u002Ffetch fails on JS-heavy pages, or when you need clean markdown\u002Fstructured extraction from a site. Requires FIRECRAWL_API_KEY.","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Fweb-scrape\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":57,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":18},"01M7EWYBN5P6F2AJG3FR24T9BP","design-taste-frontend","tasteskill v2, the anti-slop frontend skill. Art direction for websites, distilled from a study of 247 curated gallery sites (One Page Love, Awwwards Sites of the Day). Deals researched directions (idea, composition, type, palette, photographs, section blueprints) from a library, enforces a strict no-slop house style with a real-browser inspector, and runs a look-and-fix loop on screenshots. Use this whenever you build, design, redesign, restyle, polish or review a landing page, marketing site, portfolio, product page, one-pager, hero or any public-facing web page, in any stack (HTML\u002FCSS, React, Next.js, Vue, Svelte, Astro, Tailwind) — also when the user says a page looks generic, templated, \"AI-made\", sloppy, or asks to make it look designed, premium, award-level or less like every other site. Not for dashboards or app screens.","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Ftaste-design\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Ftaste-design\u002FSKILL.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":18},"01M7EWY8FQFNQ2GB7NZZWGSSDX","test-driven-development","Use when implementing any feature or bugfix, before writing implementation code","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Fsuperpowers-tdd\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":18},"01M7EWY5SHBFRH1HDHVWM9K7TV","systematic-debugging","Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Fsuperpowers-debugging\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":18},"01M7EWY3NM0Z0JAKS7HDZ24K66","requesting-code-review","Use when completing tasks, implementing major features, or before merging to verify work meets requirements","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Fsuperpowers-code-review\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":18},"01M7EWY0WRMWX4M9ZFY0WKC593","brainstorming","You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).","https:\u002F\u002Fgithub.com\u002FTanNhatCMS\u002Fjarvis\u002Fblob\u002Fmain\u002Fskills\u002Fsuperpowers-brainstorming\u002FSKILL.md",[],[]]