[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6NQ6VJVPM37YA73T8BCD3VF":3,"skill-same:01M6NQ6VJVPM37YA73T8BCD3VF":253},{"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":34,"publishedAt":55,"linkOnly":56,"markdownContent":57,"markdownBody":58,"frontMatter":59,"refName":62,"commitSha":63,"skillDirectory":64,"forksCount":14,"repositoryHtmlUrl":65,"repositoryHomepageUrl":6,"ownerHtmlUrl":66,"ownerBlogUrl":67,"ownerDescription":6,"updatedAt":68,"install":69,"resources":82},"01M6NQ6VJVPM37YA73T8BCD3VF","playwright-ui-locator",null,"Auto-heal Playwright tests after a frontend developer changes a div's class, id, or other selector. Catches a failing Playwright test, captures the old vs new DOM tree, finds the element that drifted, derives the most stable replacement locator, and proposes the fix as a GitHub pull request (confident heals) or files a GitHub issue (ambiguous \u002F element-removed cases) — never editing the working tree directly. Trigger when Playwright tests fail with \"locator resolved to 0 elements\", \"Timeout waiting for selector\", or after a UI refactor \u002F class rename \u002F Tailwind cleanup broke previously-passing tests.","carlosmarte\u002Fai-agent-artifacts","carlosmarte","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F3478576?v=4","01M6NPRTEXZG86XJXMGQ05A24X","Carlos Marte","01M6NPRTK4A6A4Q8CMWMH1ZY7W",0,"MIT","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002FSKILL.md",[18,23,26,29],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":27,"categoryName":28,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000009","前端开发",{"categoryNo":30,"categoryName":31,"taxonomyType":32,"taxonomyTypeDesc":33},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54],"playwright","playwright ui locator","ui","auto","frontend","request","element","false","rename","files","issue","never","cleanup","snapshot","fix","edit","write","read","Bash","Tailwind","2026-11-10 10:12:30",false,"---\nname: playwright-ui-locator\ndescription: Auto-heal Playwright tests after a frontend developer changes a div's class, id, or other selector. Catches a failing Playwright test, captures the old vs new DOM tree, finds the element that drifted, derives the most stable replacement locator, and proposes the fix as a GitHub pull request (confident heals) or files a GitHub issue (ambiguous \u002F element-removed cases) — never editing the working tree directly. Trigger when Playwright tests fail with \"locator resolved to 0 elements\", \"Timeout waiting for selector\", or after a UI refactor \u002F class rename \u002F Tailwind cleanup broke previously-passing tests.\nallowed-tools: Bash,Read,Write,Edit,Glob,Grep\nargument-hint: \"\u003Cfailing-test-file> [--report \u003Cplaywright-json-report>] [--baseline \u003Cold-dom-snapshot>] [--repo \u003Cowner\u002Frepo>]\"\ndisable-model-invocation: false\n---\n\n# Playwright UI Locator Auto-Healer\n\nRestore a broken Playwright test after a CSS class, `id`, or DOM-shape change has invalidated its locator. The skill orchestrates: detect the failing selector → rebuild the old element fingerprint → match it against the new DOM → pick the most stable replacement → **deliver the fix through GitHub for human review**.\n\n> **The skill never edits the user's working tree or current branch.** Confident heals are\n> applied on a fresh branch and proposed as a **pull request**; ambiguous \u002F element-removed\n> cases are filed as a **GitHub issue** with the candidate scores. All GitHub interaction goes\n> through the official `gh` CLI — see [GitHub delivery](#github-delivery-pr-or-issue) below.\n\nThe heavy lifting lives in the bundled TypeScript package at [`scripts\u002Fpackages\u002Fts\u002F`](scripts\u002Fpackages\u002Fts\u002F). The agent's job in this skill is to drive that package against the user's repo, not to re-implement DOM diffing in prompt-land.\n\n## When to use\n\n- A Playwright test that was green yesterday now errors with `locator.click: Timeout`, `strict mode violation`, or `expected count: 1, received: 0`.\n- A frontend dev renamed Tailwind classes, swapped `id` for `data-testid`, or migrated from `div > span` to `button > span`.\n- A design-system upgrade replaced raw element types (`\u003Cdiv role=\"button\">` → `\u003Cbutton>`).\n- The user asks to \"fix the broken Playwright tests\" or \"auto-update the selectors\".\n\nDo NOT trigger for:\n\n- New tests being written from scratch (use `playwright-feature-validator`).\n- Failures unrelated to selectors (assertion-value drift, backend changes, flaky waits).\n\n## Inputs\n\nThe user provides at least one of:\n\n1. A path to a failing test file: `tests\u002Fcheckout.spec.ts`\n2. A Playwright JSON report (preferred): produced by `npx playwright test --reporter=json > report.json`\n3. An optional baseline DOM snapshot (HTML file, or path to a previous `page.content()` dump). If absent, the package falls back to git history — it `git show HEAD~1:\u003Crendered-fixture>` if a fixture exists, otherwise it diffs against the locator string itself.\n\nIf only a directory is given, scan it for `*.spec.ts` \u002F `*.spec.js` files and ask which one to heal.\n\n## Step-by-step procedure\n\n### 1. Verify the failure\n\nBefore touching anything, confirm the test actually fails for a locator reason — not for a logic bug. Run:\n\n```bash\ncd \u003Crepo-root>\nnpx playwright test \u003Ctest-file> --reporter=json > \u002Ftmp\u002Fpw-report.json || true\nnode scripts\u002Fpackages\u002Fts\u002Fdist\u002Fcli.js classify --report \u002Ftmp\u002Fpw-report.json\n```\n\nOutput is one of: `LOCATOR_DRIFT` (proceed), `ASSERTION_DRIFT` (stop — not this skill's job), `INFRA_ERROR` (stop — environment issue), `UNKNOWN` (ask the user).\n\n### 2. Capture the new DOM\n\nRe-run the failing test with tracing on so the package can read the rendered page state at the moment of failure:\n\n```bash\nnpx playwright test \u003Ctest-file> --trace on || true\nnode scripts\u002Fpackages\u002Fts\u002Fdist\u002Fcli.js capture \\\n  --trace test-results\u002F\u003C...>\u002Ftrace.zip \\\n  --out \u002Ftmp\u002Fpw-newdom.html\n```\n\nIf the test fails _before_ navigation completes, the package falls back to spawning the same URL via a one-shot headless run and dumping `page.content()`.\n\n### 3. Build the old-element fingerprint\n\nFor each failing locator string in the report, the package builds a fingerprint:\n\n- Tag name\n- Visible text content (trimmed, collapsed whitespace)\n- `role`, `aria-label`, `data-testid`, `name`, `placeholder`, `title`\n- Sibling index within its parent\n- Ancestor chain (up to 3 levels) — tag + role\n\nSources, in priority order: (a) explicit `--baseline` snapshot, (b) git-history-resolved fixture, (c) the locator string itself parsed into a partial fingerprint.\n\n### 4. Score candidates in the new DOM\n\nEvery visible element in the captured new-DOM is scored against the fingerprint. Weights (see [`references\u002Fscoring.md`](references\u002Fscoring.md) for the full table):\n\n| Signal              | Weight |\n| ------------------- | ------ |\n| `data-testid` match | 100    |\n| Text content match  | 60     |\n| `role` + `name`     | 50     |\n| Tag + ancestor path | 30     |\n| `aria-label`        | 25     |\n| `id`                | 20     |\n| Class overlap       | 10     |\n\nThe highest-scoring candidate above a confidence threshold (default `0.75`) wins. Below threshold → ask the user.\n\n### 5. Derive a stable replacement locator\n\nPrefer locators in this order — pick the first that uniquely identifies the candidate:\n\n1. `getByTestId('…')` — if `data-testid` present\n2. `getByRole('…', { name: '…' })` — semantic + accessible name\n3. `getByLabel('…')` \u002F `getByPlaceholder('…')` — form fields\n4. `getByText('…', { exact: true })` — static labels\n5. `locator('#id')` — stable id\n6. `locator('[data-…]')` — other stable data attributes\n7. `locator('css-path')` — last resort, only if no semantic anchor exists\n\nClass selectors (`.btn-primary`) are _never_ emitted — they're the failure mode the skill exists to escape.\n\n### 6. Produce a delivery plan (no working-tree writes)\n\nRun the heal in **analysis-only** mode and turn the result into a GitHub delivery plan. This step writes *artifacts* (PR\u002Fissue bodies + a manifest) to an output dir — it does **not** touch git, the working tree, or the network.\n\n```bash\nnode scripts\u002Fpackages\u002Fts\u002Fdist\u002Fcli.js github-plan \\\n  --report \u002Ftmp\u002Fpw-report.json \\\n  --new-dom \u002Ftmp\u002Fpw-newdom.html \\\n  --baseline \u002Ftmp\u002Fold-dom.html \\\n  --threshold 0.8 \\\n  --out .\u002F.pw-heal-plan \\\n  --repo \u003Cowner\u002Frepo>          # optional; falls back to the origin remote\n```\n\nInternally this runs a **dry-run** heal (the AST rewriter computes the diff but writes nothing) and splits the results:\n\n- **`HEALED`** (above threshold, unambiguous) → a single aggregated **pull request**, with each drifted locator replaced on a dedicated branch.\n- **`AMBIGUOUS` \u002F `ELEMENT_REMOVED` \u002F `HEAL_FAILED`** → one **GitHub issue** each, documenting the drift, the Playwright error, and the ranked candidate scores so a human can decide.\n\nClass selectors (`.btn-primary`) are never emitted — they are the failure mode the skill exists to escape. A replacement that would match >1 element is downgraded to an issue rather than risking a strict-mode failure.\n\n### 7. Review the proposed files, then deliver\n\n**Always show the user the plan first** (the PR diff\u002Fbody and any issue bodies) and get a go-ahead before anything is pushed. The driver has a built-in review gate:\n\n```bash\n# Review gate — prints the PR body, branch, and issue titles; makes NO changes.\nscripts\u002Fopen-github.sh --plan .\u002F.pw-heal-plan --plan-only\n```\n\nOnce the user approves, deliver via the **`gh` CLI** (never the `github-create_pull_request` MCP tool):\n\n```bash\nscripts\u002Fopen-github.sh --plan .\u002F.pw-heal-plan --repo-dir \u003Crepo> [--base main] [--draft]\n```\n\nThe driver, for the PR path: refuses to run on a dirty tree, branches off the base, replays each rewrite on that branch, commits, pushes, and opens the PR with `gh pr create` — then switches back to the original branch so the user's checkout is untouched. For the issue path it runs `gh issue create` per finding.\n\nIt honors a custom GitHub endpoint and token via `GITHUB_TOKEN` and `GITHUB_BASE_API` (see [GitHub delivery](#github-delivery-pr-or-issue)).\n\n## Output\n\nEnd the turn with a structured report:\n\n```\n## playwright-ui-locator result\n\nTest:        tests\u002Fcheckout.spec.ts:42\nStatus:      HEALED → PR opened\nConfidence:  0.91\n\nOld locator: page.locator('.btn-primary')\nNew locator: page.getByRole('button', { name: 'Place order' })\n\nDelivery:    PR  #128  playwright-heal\u002F20260526-013337 → main\n             https:\u002F\u002Fgithub.com\u002Facme\u002Fstorefront\u002Fpull\u002F128\n\nAlso filed:  Issue #129 — tests\u002Fcart.spec.ts:11 (AMBIGUOUS)\n```\n\nIf nothing could be confidently healed, report the filed issues instead of a PR. The working tree is reported as **unchanged** in every case.\n\n## GitHub delivery (PR or issue)\n\nThis skill **does not modify code in place.** Every fix is delivered through GitHub for review:\n\n- **Tooling:** only the official `gh` CLI is used. The MCP `github-create_pull_request` tool is **never** invoked, and the driver does not fall back to it — `gh` gives complete, native GitHub coverage. Verify the active repo with `git status` \u002F `git branch` before delivering.\n- **PR for confident heals, issue for the rest** — the split is decided by `github-plan`; see the matrix in [`references\u002Fgithub-delivery.md`](references\u002Fgithub-delivery.md).\n- **Custom endpoint & token:**\n  - `GITHUB_TOKEN` — auth token. Exported to `GH_TOKEN` (and `GH_ENTERPRISE_TOKEN` when an enterprise host is in play). `gh` also reads `GITHUB_TOKEN` natively.\n  - `GITHUB_BASE_API` — a custom API base such as `https:\u002F\u002Fgithub.example.com\u002Fapi\u002Fv3`. Its host is derived and exported as `GH_HOST` so `gh` targets the right server.\n- **No silent edits:** the user's current branch and working tree are never mutated. Confident fixes live only on the heal branch \u002F in the PR.\n\nFull procedure, env handling, and the gh-enforcement rationale: [`references\u002Fgithub-delivery.md`](references\u002Fgithub-delivery.md).\n\n## Edge cases\n\n- **Multiple failing locators in one test** — confidently-healed ones are aggregated into a single PR; any that can't be resolved are filed as separate issues. A drift never blocks the others.\n- **The element was actually removed** — score \u003C threshold for every candidate; status `ELEMENT_REMOVED` → filed as an issue, never a PR.\n- **The test depends on visual layout** (`.screenshot()` assertions) — heal the locator in the PR, and note in the PR body that the snapshot baseline may need refreshing separately.\n- **Strict-mode duplicate match** — a replacement that would match >1 element is downgraded to an issue (with the candidate table) rather than emitting a risky `.nth()` \u002F `.filter({ hasText: … })` locator.\n- **Repo uses Playwright fixtures \u002F POM objects** — if the locator lives in a Page Object file, the PR rewrites the POM, not the spec (the failing-locator record carries the owning file path).\n\n## Test structure reference\n\nTo know *where* a drifted locator lives — inline spec, worker-scoped auth fixture, or Page Object — see [`references\u002Ftest-structure.md`](references\u002Ftest-structure.md). It documents the canonical suite layouts (minimal config + `tests\u002F`, login-once-per-worker via `storageState`, screenshot-on-failure, POM) adapted from the upstream [`microsoft\u002Fplaywright-test\u002Fexamples`](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fplaywright-test\u002Ftree\u002Fmaster\u002Fexamples), with the modern `@playwright\u002Ftest` equivalents and a per-pattern note on which file the healer should rewrite.\n\n## TypeScript package Initial Setup\n\nThe implementation lives at [`scripts\u002Fpackages\u002Fts\u002F`](scripts\u002Fpackages\u002Fts\u002F). Build it once with the setup script (resolves its own path, so it works from any directory):\n\n```bash\n.\u002Fsetup.sh            # install deps + build\n.\u002Fsetup.sh --clean    # wipe dist\u002F + node_modules first, then rebuild\n.\u002Fsetup.sh --check    # also verify the built CLI is runnable\n```\n\nEquivalent manual steps:\n\n```bash\ncd scripts\u002Fpackages\u002Fts && npm install && npm run build\n```\n\nThen the `cli.js` entry exposes: `classify`, `capture`, `fingerprint`, `match`, `rewrite`, `heal` (orchestrates the analysis), and `github-plan` (turns analysis into a PR\u002Fissue delivery plan). The `scripts\u002Fopen-github.sh` driver consumes a plan and runs the `gh` commands.\n\nSee [`scripts\u002Fpackages\u002Fts\u002FREADME.md`](scripts\u002Fpackages\u002Fts\u002FREADME.md) for the full subcommand reference and [`references\u002Fgithub-delivery.md`](references\u002Fgithub-delivery.md) for the delivery flow.\n\n## Testing the skill\n\nA two-tier harness lives at [`.test-harness\u002F`](.test-harness\u002F) with labeled mock-DOM scenarios (class rename → PR, role\u002Ftext rename → PR, element removed → issue, ambiguous duplicate → issue):\n\n```bash\ncd .test-harness\nmake fixtures   # deterministic: drives the CLI, scores vs ground truth (no API key)\nmake claude     # agentic: drives the real skill via the Claude Agent SDK (needs ANTHROPIC_API_KEY)\n```\n\nBoth emit accuracy \u002F PR-decision precision-recall metrics; the agentic tier also reports cost, turns, and skill-usage rate. See [`.test-harness\u002FREADME.md`](.test-harness\u002FREADME.md).\n","# Playwright UI Locator Auto-Healer\n\nRestore a broken Playwright test after a CSS class, `id`, or DOM-shape change has invalidated its locator. The skill orchestrates: detect the failing selector → rebuild the old element fingerprint → match it against the new DOM → pick the most stable replacement → **deliver the fix through GitHub for human review**.\n\n> **The skill never edits the user's working tree or current branch.** Confident heals are\n> applied on a fresh branch and proposed as a **pull request**; ambiguous \u002F element-removed\n> cases are filed as a **GitHub issue** with the candidate scores. All GitHub interaction goes\n> through the official `gh` CLI — see [GitHub delivery](#github-delivery-pr-or-issue) below.\n\nThe heavy lifting lives in the bundled TypeScript package at [`scripts\u002Fpackages\u002Fts\u002F`](scripts\u002Fpackages\u002Fts\u002F). The agent's job in this skill is to drive that package against the user's repo, not to re-implement DOM diffing in prompt-land.\n\n## When to use\n\n- A Playwright test that was green yesterday now errors with `locator.click: Timeout`, `strict mode violation`, or `expected count: 1, received: 0`.\n- A frontend dev renamed Tailwind classes, swapped `id` for `data-testid`, or migrated from `div > span` to `button > span`.\n- A design-system upgrade replaced raw element types (`\u003Cdiv role=\"button\">` → `\u003Cbutton>`).\n- The user asks to \"fix the broken Playwright tests\" or \"auto-update the selectors\".\n\nDo NOT trigger for:\n\n- New tests being written from scratch (use `playwright-feature-validator`).\n- Failures unrelated to selectors (assertion-value drift, backend changes, flaky waits).\n\n## Inputs\n\nThe user provides at least one of:\n\n1. A path to a failing test file: `tests\u002Fcheckout.spec.ts`\n2. A Playwright JSON report (preferred): produced by `npx playwright test --reporter=json > report.json`\n3. An optional baseline DOM snapshot (HTML file, or path to a previous `page.content()` dump). If absent, the package falls back to git history — it `git show HEAD~1:\u003Crendered-fixture>` if a fixture exists, otherwise it diffs against the locator string itself.\n\nIf only a directory is given, scan it for `*.spec.ts` \u002F `*.spec.js` files and ask which one to heal.\n\n## Step-by-step procedure\n\n### 1. Verify the failure\n\nBefore touching anything, confirm the test actually fails for a locator reason — not for a logic bug. Run:\n\n```bash\ncd \u003Crepo-root>\nnpx playwright test \u003Ctest-file> --reporter=json > \u002Ftmp\u002Fpw-report.json || true\nnode scripts\u002Fpackages\u002Fts\u002Fdist\u002Fcli.js classify --report \u002Ftmp\u002Fpw-report.json\n```\n\nOutput is one of: `LOCATOR_DRIFT` (proceed), `ASSERTION_DRIFT` (stop — not this skill's job), `INFRA_ERROR` (stop — environment issue), `UNKNOWN` (ask the user).\n\n### 2. Capture the new DOM\n\nRe-run the failing test with tracing on so the package can read the rendered page state at the moment of failure:\n\n```bash\nnpx playwright test \u003Ctest-file> --trace on || true\nnode scripts\u002Fpackages\u002Fts\u002Fdist\u002Fcli.js capture \\\n  --trace test-results\u002F\u003C...>\u002Ftrace.zip \\\n  --out \u002Ftmp\u002Fpw-newdom.html\n```\n\nIf the test fails _before_ navigation completes, the package falls back to spawning the same URL via a one-shot headless run and dumping `page.content()`.\n\n### 3. Build the old-element fingerprint\n\nFor each failing locator string in the report, the package builds a fingerprint:\n\n- Tag name\n- Visible text content (trimmed, collapsed whitespace)\n- `role`, `aria-label`, `data-testid`, `name`, `placeholder`, `title`\n- Sibling index within its parent\n- Ancestor chain (up to 3 levels) — tag + role\n\nSources, in priority order: (a) explicit `--baseline` snapshot, (b) git-history-resolved fixture, (c) the locator string itself parsed into a partial fingerprint.\n\n### 4. Score candidates in the new DOM\n\nEvery visible element in the captured new-DOM is scored against the fingerprint. Weights (see [`references\u002Fscoring.md`](references\u002Fscoring.md) for the full table):\n\n| Signal              | Weight |\n| ------------------- | ------ |\n| `data-testid` match | 100    |\n| Text content match  | 60     |\n| `role` + `name`     | 50     |\n| Tag + ancestor path | 30     |\n| `aria-label`        | 25     |\n| `id`                | 20     |\n| Class overlap       | 10     |\n\nThe highest-scoring candidate above a confidence threshold (default `0.75`) wins. Below threshold → ask the user.\n\n### 5. Derive a stable replacement locator\n\nPrefer locators in this order — pick the first that uniquely identifies the candidate:\n\n1. `getByTestId('…')` — if `data-testid` present\n2. `getByRole('…', { name: '…' })` — semantic + accessible name\n3. `getByLabel('…')` \u002F `getByPlaceholder('…')` — form fields\n4. `getByText('…', { exact: true })` — static labels\n5. `locator('#id')` — stable id\n6. `locator('[data-…]')` — other stable data attributes\n7. `locator('css-path')` — last resort, only if no semantic anchor exists\n\nClass selectors (`.btn-primary`) are _never_ emitted — they're the failure mode the skill exists to escape.\n\n### 6. Produce a delivery plan (no working-tree writes)\n\nRun the heal in **analysis-only** mode and turn the result into a GitHub delivery plan. This step writes *artifacts* (PR\u002Fissue bodies + a manifest) to an output dir — it does **not** touch git, the working tree, or the network.\n\n```bash\nnode scripts\u002Fpackages\u002Fts\u002Fdist\u002Fcli.js github-plan \\\n  --report \u002Ftmp\u002Fpw-report.json \\\n  --new-dom \u002Ftmp\u002Fpw-newdom.html \\\n  --baseline \u002Ftmp\u002Fold-dom.html \\\n  --threshold 0.8 \\\n  --out .\u002F.pw-heal-plan \\\n  --repo \u003Cowner\u002Frepo>          # optional; falls back to the origin remote\n```\n\nInternally this runs a **dry-run** heal (the AST rewriter computes the diff but writes nothing) and splits the results:\n\n- **`HEALED`** (above threshold, unambiguous) → a single aggregated **pull request**, with each drifted locator replaced on a dedicated branch.\n- **`AMBIGUOUS` \u002F `ELEMENT_REMOVED` \u002F `HEAL_FAILED`** → one **GitHub issue** each, documenting the drift, the Playwright error, and the ranked candidate scores so a human can decide.\n\nClass selectors (`.btn-primary`) are never emitted — they are the failure mode the skill exists to escape. A replacement that would match >1 element is downgraded to an issue rather than risking a strict-mode failure.\n\n### 7. Review the proposed files, then deliver\n\n**Always show the user the plan first** (the PR diff\u002Fbody and any issue bodies) and get a go-ahead before anything is pushed. The driver has a built-in review gate:\n\n```bash\n# Review gate — prints the PR body, branch, and issue titles; makes NO changes.\nscripts\u002Fopen-github.sh --plan .\u002F.pw-heal-plan --plan-only\n```\n\nOnce the user approves, deliver via the **`gh` CLI** (never the `github-create_pull_request` MCP tool):\n\n```bash\nscripts\u002Fopen-github.sh --plan .\u002F.pw-heal-plan --repo-dir \u003Crepo> [--base main] [--draft]\n```\n\nThe driver, for the PR path: refuses to run on a dirty tree, branches off the base, replays each rewrite on that branch, commits, pushes, and opens the PR with `gh pr create` — then switches back to the original branch so the user's checkout is untouched. For the issue path it runs `gh issue create` per finding.\n\nIt honors a custom GitHub endpoint and token via `GITHUB_TOKEN` and `GITHUB_BASE_API` (see [GitHub delivery](#github-delivery-pr-or-issue)).\n\n## Output\n\nEnd the turn with a structured report:\n\n```\n## playwright-ui-locator result\n\nTest:        tests\u002Fcheckout.spec.ts:42\nStatus:      HEALED → PR opened\nConfidence:  0.91\n\nOld locator: page.locator('.btn-primary')\nNew locator: page.getByRole('button', { name: 'Place order' })\n\nDelivery:    PR  #128  playwright-heal\u002F20260526-013337 → main\n             https:\u002F\u002Fgithub.com\u002Facme\u002Fstorefront\u002Fpull\u002F128\n\nAlso filed:  Issue #129 — tests\u002Fcart.spec.ts:11 (AMBIGUOUS)\n```\n\nIf nothing could be confidently healed, report the filed issues instead of a PR. The working tree is reported as **unchanged** in every case.\n\n## GitHub delivery (PR or issue)\n\nThis skill **does not modify code in place.** Every fix is delivered through GitHub for review:\n\n- **Tooling:** only the official `gh` CLI is used. The MCP `github-create_pull_request` tool is **never** invoked, and the driver does not fall back to it — `gh` gives complete, native GitHub coverage. Verify the active repo with `git status` \u002F `git branch` before delivering.\n- **PR for confident heals, issue for the rest** — the split is decided by `github-plan`; see the matrix in [`references\u002Fgithub-delivery.md`](references\u002Fgithub-delivery.md).\n- **Custom endpoint & token:**\n  - `GITHUB_TOKEN` — auth token. Exported to `GH_TOKEN` (and `GH_ENTERPRISE_TOKEN` when an enterprise host is in play). `gh` also reads `GITHUB_TOKEN` natively.\n  - `GITHUB_BASE_API` — a custom API base such as `https:\u002F\u002Fgithub.example.com\u002Fapi\u002Fv3`. Its host is derived and exported as `GH_HOST` so `gh` targets the right server.\n- **No silent edits:** the user's current branch and working tree are never mutated. Confident fixes live only on the heal branch \u002F in the PR.\n\nFull procedure, env handling, and the gh-enforcement rationale: [`references\u002Fgithub-delivery.md`](references\u002Fgithub-delivery.md).\n\n## Edge cases\n\n- **Multiple failing locators in one test** — confidently-healed ones are aggregated into a single PR; any that can't be resolved are filed as separate issues. A drift never blocks the others.\n- **The element was actually removed** — score \u003C threshold for every candidate; status `ELEMENT_REMOVED` → filed as an issue, never a PR.\n- **The test depends on visual layout** (`.screenshot()` assertions) — heal the locator in the PR, and note in the PR body that the snapshot baseline may need refreshing separately.\n- **Strict-mode duplicate match** — a replacement that would match >1 element is downgraded to an issue (with the candidate table) rather than emitting a risky `.nth()` \u002F `.filter({ hasText: … })` locator.\n- **Repo uses Playwright fixtures \u002F POM objects** — if the locator lives in a Page Object file, the PR rewrites the POM, not the spec (the failing-locator record carries the owning file path).\n\n## Test structure reference\n\nTo know *where* a drifted locator lives — inline spec, worker-scoped auth fixture, or Page Object — see [`references\u002Ftest-structure.md`](references\u002Ftest-structure.md). It documents the canonical suite layouts (minimal config + `tests\u002F`, login-once-per-worker via `storageState`, screenshot-on-failure, POM) adapted from the upstream [`microsoft\u002Fplaywright-test\u002Fexamples`](https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fplaywright-test\u002Ftree\u002Fmaster\u002Fexamples), with the modern `@playwright\u002Ftest` equivalents and a per-pattern note on which file the healer should rewrite.\n\n## TypeScript package Initial Setup\n\nThe implementation lives at [`scripts\u002Fpackages\u002Fts\u002F`](scripts\u002Fpackages\u002Fts\u002F). Build it once with the setup script (resolves its own path, so it works from any directory):\n\n```bash\n.\u002Fsetup.sh            # install deps + build\n.\u002Fsetup.sh --clean    # wipe dist\u002F + node_modules first, then rebuild\n.\u002Fsetup.sh --check    # also verify the built CLI is runnable\n```\n\nEquivalent manual steps:\n\n```bash\ncd scripts\u002Fpackages\u002Fts && npm install && npm run build\n```\n\nThen the `cli.js` entry exposes: `classify`, `capture`, `fingerprint`, `match`, `rewrite`, `heal` (orchestrates the analysis), and `github-plan` (turns analysis into a PR\u002Fissue delivery plan). The `scripts\u002Fopen-github.sh` driver consumes a plan and runs the `gh` commands.\n\nSee [`scripts\u002Fpackages\u002Fts\u002FREADME.md`](scripts\u002Fpackages\u002Fts\u002FREADME.md) for the full subcommand reference and [`references\u002Fgithub-delivery.md`](references\u002Fgithub-delivery.md) for the delivery flow.\n\n## Testing the skill\n\nA two-tier harness lives at [`.test-harness\u002F`](.test-harness\u002F) with labeled mock-DOM scenarios (class rename → PR, role\u002Ftext rename → PR, element removed → issue, ambiguous duplicate → issue):\n\n```bash\ncd .test-harness\nmake fixtures   # deterministic: drives the CLI, scores vs ground truth (no API key)\nmake claude     # agentic: drives the real skill via the Claude Agent SDK (needs ANTHROPIC_API_KEY)\n```\n\nBoth emit accuracy \u002F PR-decision precision-recall metrics; the agentic tier also reports cost, turns, and skill-usage rate. See [`.test-harness\u002FREADME.md`](.test-harness\u002FREADME.md).\n",{"name":5,"description":7,"allowed-tools":60,"argument-hint":61,"disable-model-invocation":56},"Bash,Read,Write,Edit,Glob,Grep","\u003Cfailing-test-file> [--report \u003Cplaywright-json-report>] [--baseline \u003Cold-dom-snapshot>] [--repo \u003Cowner\u002Frepo>]","main","b78f8878258c912b66be9d4b56b316515e44c728",".claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts","https:\u002F\u002Fgithub.com\u002Fcarlosmarte","http:\u002F\u002Fcarlosmarte.me\u002F","2026-08-25 00:53:46",{"source":8,"skillName":5,"commands":70,"prompt":80,"zipAvailable":81},[71,74,77],{"tool":72,"command":73},"npx","npx skills add carlosmarte\u002Fai-agent-artifacts --skill playwright-ui-locator",{"tool":75,"command":76},"bunx","bunx skills add carlosmarte\u002Fai-agent-artifacts --skill playwright-ui-locator",{"tool":78,"command":79},"pnpm","pnpm dlx skills add carlosmarte\u002Fai-agent-artifacts --skill playwright-ui-locator","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002FSKILL.md",true,[83,89,95,99,105,109,113,117,121,125,129,133,137,141,145,149,153,157,161,165,169,173,177,181,185,189,193,197,201,205,209,213,217,221,225,229,233,237,241,245,249],{"path":84,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":87,"sourceHtmlUrl":88},".test-harness\u002F.gitignore","OTHER","其他",29,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002F.gitignore",{"path":90,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":93,"sourceHtmlUrl":94},".test-harness\u002Fclaude\u002Fpackage-lock.json","REFERENCE","参考文档",48308,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Fclaude\u002Fpackage-lock.json",{"path":96,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":97,"sourceHtmlUrl":98},".test-harness\u002Fclaude\u002Fpackage.json",420,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Fclaude\u002Fpackage.json",{"path":100,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":103,"sourceHtmlUrl":104},".test-harness\u002Fclaude\u002Frun-skill.mjs","SCRIPT","脚本",12313,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Fclaude\u002Frun-skill.mjs",{"path":106,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":107,"sourceHtmlUrl":108},".test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fbaseline.html",256,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fbaseline.html",{"path":110,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":111,"sourceHtmlUrl":112},".test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fnew.html",268,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fnew.html",{"path":114,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":115,"sourceHtmlUrl":116},".test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fscenario.json",659,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fscenario.json",{"path":118,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":119,"sourceHtmlUrl":120},".test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fspec.ts",243,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fa-testid-class-rename\u002Fspec.ts",{"path":122,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":123,"sourceHtmlUrl":124},".test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fbaseline.html",146,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fbaseline.html",{"path":126,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":127,"sourceHtmlUrl":128},".test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fnew.html",155,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fnew.html",{"path":130,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":131,"sourceHtmlUrl":132},".test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fscenario.json",735,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fscenario.json",{"path":134,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":135,"sourceHtmlUrl":136},".test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fspec.ts",208,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fb-role-text-rename\u002Fspec.ts",{"path":138,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":139,"sourceHtmlUrl":140},".test-harness\u002Ffixtures\u002Fc-element-removed\u002Fbaseline.html",160,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fc-element-removed\u002Fbaseline.html",{"path":142,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":143,"sourceHtmlUrl":144},".test-harness\u002Ffixtures\u002Fc-element-removed\u002Fnew.html",176,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fc-element-removed\u002Fnew.html",{"path":146,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":147,"sourceHtmlUrl":148},".test-harness\u002Ffixtures\u002Fc-element-removed\u002Fscenario.json",642,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fc-element-removed\u002Fscenario.json",{"path":150,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":151,"sourceHtmlUrl":152},".test-harness\u002Ffixtures\u002Fc-element-removed\u002Fspec.ts",242,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fc-element-removed\u002Fspec.ts",{"path":154,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":155,"sourceHtmlUrl":156},".test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fbaseline.html",178,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fbaseline.html",{"path":158,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":159,"sourceHtmlUrl":160},".test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fnew.html",283,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fnew.html",{"path":162,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":163,"sourceHtmlUrl":164},".test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fscenario.json",641,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fscenario.json",{"path":166,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":167,"sourceHtmlUrl":168},".test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fspec.ts",253,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Ffixtures\u002Fd-ambiguous-duplicate\u002Fspec.ts",{"path":170,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":171,"sourceHtmlUrl":172},".test-harness\u002FMakefile",1126,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002FMakefile",{"path":174,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":175,"sourceHtmlUrl":176},".test-harness\u002FREADME.md",5951,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002FREADME.md",{"path":178,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":179,"sourceHtmlUrl":180},".test-harness\u002Frun-fixtures.mjs",8614,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002F.test-harness\u002Frun-fixtures.mjs",{"path":182,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":183,"sourceHtmlUrl":184},"references\u002Fgithub-delivery.md",5657,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Freferences\u002Fgithub-delivery.md",{"path":186,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":187,"sourceHtmlUrl":188},"references\u002Fscoring.md",3312,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Freferences\u002Fscoring.md",{"path":190,"resourceType":91,"resourceTypeDesc":92,"sizeBytes":191,"sourceHtmlUrl":192},"references\u002Ftest-structure.md",5269,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Freferences\u002Ftest-structure.md",{"path":194,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":195,"sourceHtmlUrl":196},"scripts\u002Fopen-github.sh",9839,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fopen-github.sh",{"path":198,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":199,"sourceHtmlUrl":200},"scripts\u002Fpackages\u002Fts\u002F.gitignore",17,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002F.gitignore",{"path":202,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":203,"sourceHtmlUrl":204},"scripts\u002Fpackages\u002Fts\u002Fpackage-lock.json",32830,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fpackage-lock.json",{"path":206,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":207,"sourceHtmlUrl":208},"scripts\u002Fpackages\u002Fts\u002Fpackage.json",787,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fpackage.json",{"path":210,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":211,"sourceHtmlUrl":212},"scripts\u002Fpackages\u002Fts\u002FREADME.md",4191,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002FREADME.md",{"path":214,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":215,"sourceHtmlUrl":216},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Fcli.ts",7453,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Fcli.ts",{"path":218,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":219,"sourceHtmlUrl":220},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Ffingerprint.ts",2158,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Ffingerprint.ts",{"path":222,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":223,"sourceHtmlUrl":224},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Fgithub-plan.ts",8881,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Fgithub-plan.ts",{"path":226,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":227,"sourceHtmlUrl":228},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Findex.ts",2981,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Findex.ts",{"path":230,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":231,"sourceHtmlUrl":232},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Flocator-resolver.ts",5714,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Flocator-resolver.ts",{"path":234,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":235,"sourceHtmlUrl":236},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Fparse-failure.ts",2989,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Fparse-failure.ts",{"path":238,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":239,"sourceHtmlUrl":240},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Frewrite-test.ts",3561,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Frewrite-test.ts",{"path":242,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":243,"sourceHtmlUrl":244},"scripts\u002Fpackages\u002Fts\u002Fsrc\u002Ftypes.ts",1236,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Fsrc\u002Ftypes.ts",{"path":246,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":247,"sourceHtmlUrl":248},"scripts\u002Fpackages\u002Fts\u002Ftsconfig.json",507,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fscripts\u002Fpackages\u002Fts\u002Ftsconfig.json",{"path":250,"resourceType":101,"resourceTypeDesc":102,"sizeBytes":251,"sourceHtmlUrl":252},"setup.sh",1875,"https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002Fsetup.sh",[254,281,305,333,356,383],{"skillNo":255,"name":256,"nameZh":6,"description":257,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":258,"categories":259,"keywords":260,"publishedAt":55},"01M6NQ7MN7FKWQE3CKRWKVAC9K","validation_handoff","Forces the agent to conclude its response by generating a strict validation checklist and outlining pending actions, focusing heavily on defensive execution and edge-case testing.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fvalidation_handoff\u002FSKILL.md",[],[261,262,263,264,265,266,267,268,269,270,271,272,273,274,275,276,277,278,279,280],"handoff","validation","validation handoff","testing","architecture","edge","actions","refactor","execution","system","implementation","checklist","feature implementation","feature","end","case","defensive","hook","at","pending",{"skillNo":282,"name":283,"nameZh":6,"description":284,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":285,"categories":286,"keywords":287,"publishedAt":55},"01M6NQ74C8EKMSABDNN4T6TFBD","project-ployglot-project-scafford","Scaffold a polyglot library project as TypeScript + Python twin packages under `packages\u002Fts\u002F` + `packages\u002Fpy\u002F`, with a shared `SPEC.md`, cross-language `tests\u002Fparity\u002Ffixtures.json`, side-by-side `examples\u002Fsdk\u002F` + `examples\u002Fapi\u002F` documentation, per-package `Makefile`, and a root orchestrator that exposes `make ci`. Open-source friendly — no organization-specific paths or scopes baked in. Use when starting a new polyglot library, porting a single-language library to a twin, or auditing\u002Fregularizing an existing two-language repo against the canonical layout.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fproject-ployglot-project-scafford\u002FSKILL.md",[],[288,289,290,291,292,293,294,295,296,53,52,297,298,299,300,301,302,42,303,304],"project","project ployglot project scafford","documentation","packages","run","ci","api","open","makefile","orchestrator","only","organization","shared","layout","dry","scope","\u003Cslug>",{"skillNo":306,"name":307,"nameZh":6,"description":308,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":309,"categories":310,"keywords":314,"publishedAt":55},"01M6NQ6BCW05C6Y6VRV0RBQD1V","org-learn-context-window","Distill the current conversation into a reusable command or skill artifact. Use when the user wants to capture a workflow they just executed (create a reusable slash command) or formalize a resolution\u002Fmistake\u002Fcorrection into a reusable skill so the same class of issue never recurs.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Forg-learn-context-window\u002FSKILL.md",[311],{"categoryNo":312,"categoryName":313,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",[315,316,317,318,319,320,321,322,323,324,325,326,327,46,45,328,329,330,331,332],"learn","context","org","window","org learn context window","workflow","create","just","capture","distill","command","user","use","reusable","conversation","artifact","resolution","skill artifact",{"skillNo":334,"name":335,"nameZh":6,"description":336,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":337,"categories":338,"keywords":340,"publishedAt":55},"01M6NQ66NFPQM1PDBHP1B2G82X","org-agent-skill-creator","Scaffold and validate scope-aware agent skills (company, enterprise, org, team, project, application). Use when the user asks to create a new agent skill, generate a SKILL.md file, scaffold a skill directory, or validate that existing skill directories meet SKILL_SPEC.md (frontmatter, naming rules, structure, scope\u002Ftier).","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Forg-agent-skill-creator\u002FSKILL.md",[339],{"categoryNo":312,"categoryName":313,"taxonomyType":21,"taxonomyTypeDesc":22},[341,342,317,343,344,345,321,346,347,348,349,327,350,351,288,303,352,353,354,355],"skill creator","agent skill","agent skill creator","creator","org agent skill creator","file","Enterprise","team","Rules","structure","new agent","directories","Existing","company","Meet",{"skillNo":357,"name":358,"nameZh":6,"description":359,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":360,"categories":361,"keywords":365,"publishedAt":55},"01M6NQ5VSH5M682YMYDF9H9KFE","app-workshop-session-pdf-recap-next","Extract content from a workshop session PDF and produce a structured recap focused on (1) what was learned, (2) next steps, and (3) possible enhancements. Ingests a single PDF file path, parses session material (slides, transcripts, exercises, Q&A), and emits a Markdown report with Learning Recap, Next Steps, and Enhancement Opportunities sections with page-level traceability. Use when the user asks to \"summarize a workshop\", \"recap a training PDF\", \"pull next steps from a session deck\", or provides a workshop\u002Fseminar\u002Ftutorial PDF and wants a learn-and-act summary.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fapp-workshop-session-pdf-recap-next\u002FSKILL.md",[362],{"categoryNo":363,"categoryName":364,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000002","文档处理",[366,367,368,369,370,371,372,315,373,374,375,376,377,378,327,379,42,380,381,382],"PDF","recap","session","next","app","app workshop session pdf recap next","training","pull","produce","learning","slides","page","tutorial","deck","material","traceability","next steps",{"skillNo":384,"name":385,"nameZh":6,"description":386,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":387,"categories":388,"keywords":390,"publishedAt":55},"01M6NQ5NFNK613NR7FM8HPAJ74","app-workshop-recaps-consolidate","Consolidate multiple workshop-session recap Markdown files (typically produced by the workshop-session-pdf-recap-next skill) into a single unified report. Merges and deduplicates learnings across sessions and emits six sections — Consolidated Recap, Follow-ups, Action Items, Next Steps, Possible Enhancements, and Industry \u002F External Viewpoint. The last section supplements the attendees' material with current external perspective via web research. Use when the user has a set of per-session recap files and wants one portfolio-level summary with an outside-in lens (\"consolidate these recaps\", \"merge my workshop notes\", \"cross-session summary\", \"roll up action items\").","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fapp-workshop-recaps-consolidate\u002FSKILL.md",[389],{"categoryNo":363,"categoryName":364,"taxonomyType":21,"taxonomyTypeDesc":22},[391,370,392,366,393,394,53,395,367,396,397,44,398,369,399,42,400,382,401,402],"consolidate","app workshop recaps consolidate","notes","SET","markdown","portfolio","up","learnings","across","lens","follow-ups","websearch"]