[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7A9FGDMN13GTCAFQSNXQDEG":3,"skill-same:01M7A9FGDMN13GTCAFQSNXQDEG":68},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":45,"commitSha":46,"skillDirectory":47,"forksCount":48,"repositoryHtmlUrl":49,"repositoryHomepageUrl":6,"ownerHtmlUrl":50,"ownerBlogUrl":51,"ownerDescription":52,"updatedAt":53,"install":54,"resources":67},"01M7A9FGDMN13GTCAFQSNXQDEG","wgm",null,"Turns a rough request into working software via a governed build loop: align first, plan, then iterate one task at a time with deterministic backpressure and holdout-scenario judging.","MHassan0000\u002FSkills","MHassan0000","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F224296448?v=4","01M7A14DTC1EZCRN04582533RD","Muhammad Hassan Yousaf","01M7A14DV10KTW8NHQTJPD08X0",1,"MIT","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fantigravity skills\u002Fagentic-awesome-skills\u002Fskills\u002Fwgm\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: wgm\ndescription: \"Turns a rough request into working software via a governed build loop: align first, plan, then iterate one task at a time with deterministic backpressure and holdout-scenario judging.\"\ncategory: meta\nrisk: safe\nsource: community\nsource_repo: agent-frontier\u002Fwgm\nsource_type: official\ndate_added: \"2026-07-05\"\nauthor: agent-frontier\ntags: [build-loop, spec-driven, ralph-loop, self-improving, agentic-development, methodology]\ntools: [claude, cursor, gemini, copilot, codex]\nlicense: \"MIT\"\nlicense_source: \"https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm\u002Fblob\u002Fmain\u002FLICENSE\"\n---\n\n# wgm\n\n## Overview\n\nwgm (\"well, gosh... make\") is a portable build **methodology**, not a domain skill — a single\n`SKILL.md` protocol that any agentskills.io-compatible host loads to turn a rough request into\nworking software. It marries three ideas: a relentless alignment interview before any code is\nwritten, a Ralph-style loop (one task per iteration, a persistent plan as shared state, steered by\ndeterministic backpressure), and holdout-scenario LLM judging (scenarios the build never sees, so a\nhigh satisfaction score can't be gamed). It also runs its own internal docs-audit and\nself-improvement loop, cross-pollinating durable lessons from sibling agent-coding projects back\ninto its own protocol.\n\n## When to Use This Skill\n\n- Use when building or implementing a feature, app, or prototype from rough or ambiguous intent.\n- Use when a task benefits from a governed plan plus iterative, test-validated execution rather\n  than one-shot generation.\n- Use when you want a build to converge against acceptance criteria an LLM judge scores blind\n  (0-100), instead of trusting a single self-reported \"looks good.\"\n- Not for trivial one-file edits, pure debugging, research-only questions, or tasks that already\n  have complete, unambiguous step-by-step instructions — wgm explicitly stays out of the way there.\n\n## How It Works\n\n### Step 1: Triage\nClassify the work onto a scale-adaptive track (Quick \u002F Standard \u002F Full) so ceremony matches risk —\na one-file fix skips holdout scenarios and the docs-audit swarm; a greenfield app gets the full rig.\nThe deterministic backpressure gate itself is never skipped, only the ceremony around it.\n\n### Step 2: Grill (align)\nInterview the user one question at a time, always with a recommended answer, until the goal,\nsuccess criteria, and constraints are known — capping interrogation after ~5 questions to avoid\ntheater. Explore the codebase to self-answer before asking anything a human doesn't need to weigh\nin on.\n\n### Step 3: Plan\nProduce a project constitution, one spec per coherent slice (each with a magic moment and a demo\npath), holdout acceptance scenarios the build must never read, and `IMPLEMENTATION_PLAN.md` — the\npersistent shared state across every later iteration. Cross-check every artifact against every other\none before moving on.\n\n### Step 4: Preflight\nScore the plan's readiness 0-100 across goal clarity, observable success criteria, scenario\ncoverage, and backpressure mapping. Below the threshold, return to Grill\u002FPlan and fix the weakest\ndimension — do not start building on a shaky plan.\n\n### Step 5: Loop (build)\nRun `Analyze -> Implement -> Validate -> Review -> Record`, one task per iteration: pick the single\nmost important pending task, make the smallest change that completes it, run its deterministic\nvalidation command (green or it isn't done), judge holdout-scenario satisfaction, review the diff\nfor scope creep, then record status and any durable lesson before advancing exactly one task.\n\n### Step 6: Ship \u002F Handoff\nSummarize what shipped and how to validate it, run a mandatory four-persona docs-audit pass\n(junior\u002Fsenior\u002Fprincipal\u002FPM perspectives, consolidated into one paper-trail report), and harvest any\ndurable, cross-project lesson back into the shared skill's own ledger.\n\n## Examples\n\n### Example 1: Full lifecycle from a rough request\n\n```\nUser: \"Build a CLI todo app with add\u002Flist\u002Fcomplete commands, from scratch.\"\n```\n\nwgm states its Track (Standard), grills for the ~3-5 unknowns that actually matter, writes specs +\n`IMPLEMENTATION_PLAN.md`, scores Preflight readiness, then loops one task at a time — each task's\nown test\u002Flint\u002Fbuild command must exit 0 before it's marked done — and finally ships with a\ndocs-audit pass.\n\n### Example 2: Scoped planning only\n\n```\nUser: \"\u002Fwgm plan: add OAuth login to this existing Express API\"\n```\n\nwgm writes the specs and plan, then hard-stops at the Plan-exit gate without starting the build loop\n— useful when a human wants to review the plan before any code is touched.\n\n## Best Practices\n\n- Do let the plan be the shared state — a fresh agent should be able to resume a build from\n  `IMPLEMENTATION_PLAN.md` alone.\n- Do keep holdout scenarios genuinely hidden from the generating agent; that's what prevents a\n  judged score from being gamed.\n- Do map every acceptance criterion to a runnable, deterministic check before calling anything done.\n- Don't skip the alignment interview on ambiguous, multi-week, or security\u002FUX-critical work just to\n  move faster — misalignment discovered after building is far more expensive.\n- Don't treat a high satisfaction score as sufficient on its own — a failing deterministic check\n  always overrides it.\n\n## Limitations\n\n- wgm is a protocol, not a runtime: it has no daemon, scheduler, or bundled dashboard — it expects\n  an existing agentskills.io-compatible host to load and execute it.\n- This skill does not replace environment-specific validation, testing, or expert review.\n- Full holdout-scenario judging and the docs-audit swarm add ceremony that a genuinely trivial task\n  does not need — wgm's own Triage track exists specifically to right-size this, and the skill\n  explicitly says not to use it for one-file edits or pure debugging.\n\n## Common Pitfalls\n\n- **Problem:** Treating wgm's \"build\" mode the same as a full-lifecycle request.\n  **Solution:** `\u002Fwgm build` resumes an *existing* `IMPLEMENTATION_PLAN.md`; a bare request like\n  \"build the auth module\" (more text after \"build\") is a full-lifecycle request, not `build` mode.\n- **Problem:** Letting the agent peek at holdout scenarios while implementing.\n  **Solution:** Scenarios are read only during Validate\u002FReview, never during Implement — that's the\n  entire point of a holdout set.\n\n## Related Skills\n\n- `@grill-me` - the narrower alignment-interview primitive wgm's Grill phase is adapted from.\n- `@skill-creator` - useful for authoring\u002Fevaluating the skill itself; wgm ships its own eval\n  fixture (`evals\u002Fevals.json`) using the same eval-driven-iteration discipline.\n\n## Additional Resources\n\n- [Repository](https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm)\n- [Full protocol (`SKILL.md`)](https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm\u002Fblob\u002Fmain\u002FSKILL.md)\n- [Reference library](https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm\u002Ftree\u002Fmain\u002Freferences)\n","# wgm\n\n## Overview\n\nwgm (\"well, gosh... make\") is a portable build **methodology**, not a domain skill — a single\n`SKILL.md` protocol that any agentskills.io-compatible host loads to turn a rough request into\nworking software. It marries three ideas: a relentless alignment interview before any code is\nwritten, a Ralph-style loop (one task per iteration, a persistent plan as shared state, steered by\ndeterministic backpressure), and holdout-scenario LLM judging (scenarios the build never sees, so a\nhigh satisfaction score can't be gamed). It also runs its own internal docs-audit and\nself-improvement loop, cross-pollinating durable lessons from sibling agent-coding projects back\ninto its own protocol.\n\n## When to Use This Skill\n\n- Use when building or implementing a feature, app, or prototype from rough or ambiguous intent.\n- Use when a task benefits from a governed plan plus iterative, test-validated execution rather\n  than one-shot generation.\n- Use when you want a build to converge against acceptance criteria an LLM judge scores blind\n  (0-100), instead of trusting a single self-reported \"looks good.\"\n- Not for trivial one-file edits, pure debugging, research-only questions, or tasks that already\n  have complete, unambiguous step-by-step instructions — wgm explicitly stays out of the way there.\n\n## How It Works\n\n### Step 1: Triage\nClassify the work onto a scale-adaptive track (Quick \u002F Standard \u002F Full) so ceremony matches risk —\na one-file fix skips holdout scenarios and the docs-audit swarm; a greenfield app gets the full rig.\nThe deterministic backpressure gate itself is never skipped, only the ceremony around it.\n\n### Step 2: Grill (align)\nInterview the user one question at a time, always with a recommended answer, until the goal,\nsuccess criteria, and constraints are known — capping interrogation after ~5 questions to avoid\ntheater. Explore the codebase to self-answer before asking anything a human doesn't need to weigh\nin on.\n\n### Step 3: Plan\nProduce a project constitution, one spec per coherent slice (each with a magic moment and a demo\npath), holdout acceptance scenarios the build must never read, and `IMPLEMENTATION_PLAN.md` — the\npersistent shared state across every later iteration. Cross-check every artifact against every other\none before moving on.\n\n### Step 4: Preflight\nScore the plan's readiness 0-100 across goal clarity, observable success criteria, scenario\ncoverage, and backpressure mapping. Below the threshold, return to Grill\u002FPlan and fix the weakest\ndimension — do not start building on a shaky plan.\n\n### Step 5: Loop (build)\nRun `Analyze -> Implement -> Validate -> Review -> Record`, one task per iteration: pick the single\nmost important pending task, make the smallest change that completes it, run its deterministic\nvalidation command (green or it isn't done), judge holdout-scenario satisfaction, review the diff\nfor scope creep, then record status and any durable lesson before advancing exactly one task.\n\n### Step 6: Ship \u002F Handoff\nSummarize what shipped and how to validate it, run a mandatory four-persona docs-audit pass\n(junior\u002Fsenior\u002Fprincipal\u002FPM perspectives, consolidated into one paper-trail report), and harvest any\ndurable, cross-project lesson back into the shared skill's own ledger.\n\n## Examples\n\n### Example 1: Full lifecycle from a rough request\n\n```\nUser: \"Build a CLI todo app with add\u002Flist\u002Fcomplete commands, from scratch.\"\n```\n\nwgm states its Track (Standard), grills for the ~3-5 unknowns that actually matter, writes specs +\n`IMPLEMENTATION_PLAN.md`, scores Preflight readiness, then loops one task at a time — each task's\nown test\u002Flint\u002Fbuild command must exit 0 before it's marked done — and finally ships with a\ndocs-audit pass.\n\n### Example 2: Scoped planning only\n\n```\nUser: \"\u002Fwgm plan: add OAuth login to this existing Express API\"\n```\n\nwgm writes the specs and plan, then hard-stops at the Plan-exit gate without starting the build loop\n— useful when a human wants to review the plan before any code is touched.\n\n## Best Practices\n\n- Do let the plan be the shared state — a fresh agent should be able to resume a build from\n  `IMPLEMENTATION_PLAN.md` alone.\n- Do keep holdout scenarios genuinely hidden from the generating agent; that's what prevents a\n  judged score from being gamed.\n- Do map every acceptance criterion to a runnable, deterministic check before calling anything done.\n- Don't skip the alignment interview on ambiguous, multi-week, or security\u002FUX-critical work just to\n  move faster — misalignment discovered after building is far more expensive.\n- Don't treat a high satisfaction score as sufficient on its own — a failing deterministic check\n  always overrides it.\n\n## Limitations\n\n- wgm is a protocol, not a runtime: it has no daemon, scheduler, or bundled dashboard — it expects\n  an existing agentskills.io-compatible host to load and execute it.\n- This skill does not replace environment-specific validation, testing, or expert review.\n- Full holdout-scenario judging and the docs-audit swarm add ceremony that a genuinely trivial task\n  does not need — wgm's own Triage track exists specifically to right-size this, and the skill\n  explicitly says not to use it for one-file edits or pure debugging.\n\n## Common Pitfalls\n\n- **Problem:** Treating wgm's \"build\" mode the same as a full-lifecycle request.\n  **Solution:** `\u002Fwgm build` resumes an *existing* `IMPLEMENTATION_PLAN.md`; a bare request like\n  \"build the auth module\" (more text after \"build\") is a full-lifecycle request, not `build` mode.\n- **Problem:** Letting the agent peek at holdout scenarios while implementing.\n  **Solution:** Scenarios are read only during Validate\u002FReview, never during Implement — that's the\n  entire point of a holdout set.\n\n## Related Skills\n\n- `@grill-me` - the narrower alignment-interview primitive wgm's Grill phase is adapted from.\n- `@skill-creator` - useful for authoring\u002Fevaluating the skill itself; wgm ships its own eval\n  fixture (`evals\u002Fevals.json`) using the same eval-driven-iteration discipline.\n\n## Additional Resources\n\n- [Repository](https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm)\n- [Full protocol (`SKILL.md`)](https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm\u002Fblob\u002Fmain\u002FSKILL.md)\n- [Reference library](https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm\u002Ftree\u002Fmain\u002Freferences)\n",{"name":5,"risk":24,"tags":25,"tools":32,"author":38,"source":39,"license":15,"category":40,"date_added":41,"description":7,"source_repo":42,"source_type":43,"license_source":44},"safe",[26,27,28,29,30,31],"build-loop","spec-driven","ralph-loop","self-improving","agentic-development","methodology",[33,34,35,36,37],"claude","cursor","gemini","copilot","codex","agent-frontier","community","meta","2026-07-05","agent-frontier\u002Fwgm","official","https:\u002F\u002Fgithub.com\u002Fagent-frontier\u002Fwgm\u002Fblob\u002Fmain\u002FLICENSE","main","a4404a75dc81d6b7e58c5906b6aaffbcb5e84747","Claude\u002Fantigravity skills\u002Fagentic-awesome-skills\u002Fskills\u002Fwgm",0,"https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills","https:\u002F\u002Fgithub.com\u002FMHassan0000","https:\u002F\u002Fhassanyousaf.vercel.app","Full Stack Developer building web apps, SaaS, AI systems, and automations. Passionate about clean code, UI\u002FUX, backend systems, and turning ideas into products.","2026-08-07 20:42:05",{"source":8,"skillName":5,"commands":55,"prompt":65,"zipAvailable":66},[56,59,62],{"tool":57,"command":58},"npx","npx skills add MHassan0000\u002FSkills --skill wgm",{"tool":60,"command":61},"bunx","bunx skills add MHassan0000\u002FSkills --skill wgm",{"tool":63,"command":64},"pnpm","pnpm dlx skills add MHassan0000\u002FSkills --skill wgm","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fantigravity skills\u002Fagentic-awesome-skills\u002Fskills\u002Fwgm\u002FSKILL.md",true,[],[69,76,83,90,99,107],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M7AGBX58S04DX8DKKME8N311","vercel-react-best-practices","React and Next.js performance optimization guidelines from Vercel Engineering. This skill should be used when writing, reviewing, or refactoring React\u002FNext.js code to ensure optimal performance patterns. Triggers on tasks involving React components, Next.js pages, data fetching, bundle optimization, or performance improvements.","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fskills boraztunc github\u002Fvercel-react-best-practices\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M7AGBREE890608Z7YDMAKH27","stop-slop","Remove AI writing patterns from prose. Use when drafting, editing, or reviewing text to eliminate predictable AI tells and produce more human-like writing.","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fskills boraztunc github\u002Fstop-slop\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M7AGAS972A2AN7Q5EW8HEAXK","ogilvy-copywriting","Apply David Ogilvy's advertising principles when writing or reviewing copy, headlines, product descriptions, landing pages, ads, emails, or any persuasive text. Use when the goal is copy that sells, not copy that wins awards. Covers positioning, headlines, promises, brand voice, long-form copy, and visual logic. Source: How to Create Advertising That Sells (1972) + Ogilvy on Advertising.","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fskills boraztunc github\u002Fogilvy\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":94,"licenseFileUrl":95,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":19},"01M7AGAF12803G5JPVMK79AYJB","frontend-design","Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML\u002FCSS layouts, or when styling\u002Fbeautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.","Apache-2.0","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fskills boraztunc github\u002Ffrontend-design\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fskills boraztunc github\u002Ffrontend-design\u002FSKILL.md",[],[],{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":94,"licenseFileUrl":103,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":19},"01M7AG88WWVWCGJY9WA9CH7EY4","webapp-testing","Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fobra superpower skills\u002Fwebapp-testing\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fobra superpower skills\u002Fwebapp-testing\u002FSKILL.md",[],[],{"skillNo":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":94,"licenseFileUrl":111,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":19},"01M7AG7ZTNV7XQCWDQ9674FBYB","web-artifacts-builder","Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn\u002Fui). Use for complex artifacts requiring state management, routing, or shadcn\u002Fui components - not for simple single-file HTML\u002FJSX artifacts.","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fobra superpower skills\u002Fweb-artifacts-builder\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FMHassan0000\u002FSkills\u002Fblob\u002Fmain\u002FClaude\u002Fobra superpower skills\u002Fweb-artifacts-builder\u002FSKILL.md",[],[]]