[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-81532":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":16,"stars7d":15,"stars30d":15,"stars90d":16,"forks30d":16,"starsTrendScore":16,"compositeScore":17,"rankGlobal":10,"rankLanguage":10,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":21,"hasPages":19,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":23,"readmeContent":24,"aiSummary":25,"trendingCount":16,"starSnapshotCount":16,"syncStatus":26,"lastSyncTime":27,"discoverSource":28},81532,"agent-readiness","superduck-ai\u002Fagent-readiness","superduck-ai","A vendor-free implementation of the Factory droid skill — runs in any coding agent","",null,"TypeScript",27,4,26,1,0,42.7,"MIT License",false,"main",true,[],"2026-06-12 04:01:34","# agent-readiness\n\n> Chinese version: [README.zh-CN.md](README.zh-CN.md)\n\nA portable agent-readiness audit skill: scans a git repository, evaluates 82\ncriteria across 9 categories via parallel subagents, and emits a JSON\nreport, persistent score history, and an HTML dashboard.\n\n## Layout\n\n- `agent-readiness\u002F` — **the skill**; the publishable artifact\n  - `SKILL.md` — the agent prompt (5-phase audit workflow)\n  - `criteria\u002F` — 9 markdown files, one per category, read by Phase-3 subagents\n  - `bin\u002Fagent-readiness.mjs` — bundled, self-contained CLI used by Phases 4–5\n- `cli\u002F` — dev workspace that produces `agent-readiness\u002Fbin\u002Fagent-readiness.mjs`\n\n## Installing the skill\n\nCopy or symlink `agent-readiness\u002F` into your host's skills directory:\n\n```sh\ncp -r agent-readiness ~\u002F.claude\u002Fskills\u002F\n```\n\nThe skill is fully self-contained — the bundled `bin\u002Fagent-readiness.mjs` has\nno external runtime dependencies beyond Node.js ≥ 18.17.\n\n## Running the audit\n\nInside any git repository, ask the host to invoke the skill (e.g. via slash\ncommand or by referencing `SKILL.md`). The skill drives the entire 5-phase\nworkflow itself and writes its artifacts to a repo-local `.agent-readiness\u002F`\ndirectory:\n\n```text\n.agent-readiness\u002F\n├── .gitignore\n├── history\u002F\n│   ├── .gitkeep\n│   └── *.json\n└── latest\u002F\n    ├── agent-readiness-report.json\n    ├── agent-readiness-score.json\n    └── agent-readiness-dashboard.html\n```\n\nRecommended git behavior:\n\n- Commit `.agent-readiness\u002Fhistory\u002F*.json` so trend data is reviewable and durable.\n- Ignore `.agent-readiness\u002Flatest\u002F` because those files are local, disposable, and fully reproducible from the latest run plus history.\n- A nested `.agent-readiness\u002F.gitignore` works well for this and keeps the repo root `.gitignore` clean.\n- The skill bootstraps `.agent-readiness\u002F.gitignore` with `latest\u002F` and ensures `.agent-readiness\u002Fhistory\u002F.gitkeep` exists on first run.\n\n## Factory Overview Supplement\n\nThis repository maps directly to the Factory Agent Readiness reporting flow: a\nstatic audit produces a raw report, a scored summary, and a browsable HTML\ndashboard.\n\n- `agent-readiness-report.json`: raw criterion evaluations\n- `agent-readiness-score.json`: computed score, current level, and history input\n- `agent-readiness-dashboard.html`: a visual dashboard for exploring the latest snapshot\n\nThe current implementation evaluates 82 criteria across 9 categories:\n\n- Style & Validation\n- Build System\n- Testing\n- Documentation\n- Development Environment\n- Debugging & Observability\n- Security\n- Task Discovery\n- Product & Experimentation\n\nEach criterion stores more than a simple pass or fail:\n\n- status: `passed`, `failed`, `skipped`, or `missing`\n- score: `numerator \u002F denominator`\n- rationale: the evidence-backed explanation shown in the dashboard\n\nFor application-scoped criteria, a score like `3\u002F4` means 3 applications pass\nout of 4 discovered applications. For repository-scoped criteria, the common\ncase is `1\u002F1` or `0\u002F1`.\n\n### Reading the overview dashboard\n\n![Factory overview dashboard](docs\u002Fimages\u002Ffactory-overview-dashboard.png)\n\nRead the overview page in this order:\n\n- repository context: name, commit, branch, and flags like `local changes`\n- overall readiness: current level and total pass rate\n- outcome mix: counts for `Passed`, `Failed`, and `Skipped`\n- `Readiness by level`: one segment per criterion, with the marker showing current overall progress\n- `Pass Rate by Category`: a radar view of strong and weak categories\n- `Level Over Time`: trend data loaded from `.agent-readiness\u002Fhistory\u002F*.json`\n\nIn practice, this page answers three questions quickly:\n\n- What level is the repository at right now?\n- Which categories are currently the weakest?\n- Is readiness improving over time?\n\n### Reading the criteria browser\n\n![Factory criteria browser](docs\u002Fimages\u002Ffactory-criteria-browser.png)\n\nThe criteria browser is where you drill from the overall score down to a\nspecific signal.\n\nThe most useful elements are:\n\n- the search box for criterion name, id, or rationale\n- status filters: `All`, `Passed`, `Failed`, and `Skipped`\n- category filters for all 9 categories\n- the category summary line, such as `2 \u002F 12 fully passed · 43.8% average score · 13 criteria`\n\nYou can read a single criterion row like this:\n\n- `L1`, `L2`, and so on indicate the readiness level for that criterion\n- `A` means application scope, while `R` means repository scope\n- `Skippable` means the criterion may legitimately resolve to `N\u002FA`\n- `3\u002F4` means 3 of 4 applications passed\n- `N\u002FA` means skipped\n- `—` means the report has no entry for that criterion\n- `rationale` is the most important field because it explains why the result was assigned\n\nWhen prioritizing work, the highest-value first fixes are usually:\n\n- failed application-scoped criteria with a large denominator\n- Level 1 or Level 2 criteria that affect everyday developer workflow\n- repository-scoped criteria that improve all apps at once\n\n### Using the Fix flow\n\n![Factory fix modal](docs\u002Fimages\u002Ffactory-fix-modal.png)\n\nWhen a criterion is `failed` or `missing`, the dashboard exposes a `Fix`\nbutton. Clicking it opens a modal with a copyable remediation prompt.\n\nThat prompt includes:\n\n- the failing signal name, score, and description\n- the original evaluation instructions used to grade the signal\n- a focused task list: inspect the repository, implement a substantive fix, then verify it\n- quality guardrails that reject placeholder files, disabled checks, or metric-gaming changes\n- a completion requirement to open a PR and return the PR URL after code changes\n\nIn this repository, the copied prompt is intentionally written for use in a\nlocal clone, so a failed signal can turn directly into an actionable\nremediation task.\n\n## Working on the CLI\n\n```sh\ncd cli\nnpm install\nnpm test\nnpm run bundle    # → ..\u002Fagent-readiness\u002Fbin\u002Fagent-readiness.mjs\n```\n\nSee [`cli\u002FREADME.md`](cli\u002FREADME.md) for details.\n","该项目是一个可移植的代理就绪性审计工具，能够在任何编码代理中运行。它通过并行子代理扫描Git仓库，根据9个类别中的82项标准进行评估，并生成JSON报告、持久化的评分历史记录以及一个HTML仪表板。项目使用TypeScript编写，具有良好的自包含性和便携性，仅需Node.js 18.17及以上版本即可运行。适用于需要对代码仓库进行全面质量检查和持续改进的开发团队或个人开发者，帮助他们快速定位项目在风格验证、构建系统、测试、文档等多方面可能存在的问题。",2,"2026-06-11 04:05:24","CREATED_QUERY"]