[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6MCXW1VKBAMRHRPGDMC3CQQ":3,"skill-same:01M6MCXW1VKBAMRHRPGDMC3CQQ":219},{"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":31,"publishedAt":52,"linkOnly":53,"markdownContent":54,"markdownBody":55,"frontMatter":56,"refName":61,"commitSha":62,"skillDirectory":63,"forksCount":64,"repositoryHtmlUrl":65,"repositoryHomepageUrl":66,"ownerHtmlUrl":67,"ownerBlogUrl":68,"ownerDescription":69,"updatedAt":70,"install":71,"resources":84},"01M6MCXW1VKBAMRHRPGDMC3CQQ","sectionsight",null,"Sets up privacy-first website analytics that show what visitors actually read (active seconds per page section and whether it was read; scroll depth, CTA and goal clicks, dead and rage clicks, copies, prints, video plays) with no cookies. Adds a consent banner designed around the EDPB Cookie Banner Taskforce report and CNIL guidance, optional PostHog session replay only after consent, a Slack or e-mail line per notable visit, a weekly digest, AI referral and crawler counts, health alerts, enforced retention, and self-service deletion of a visitor's data. Then proves it in a real browser (nothing stored and no ids before consent, no e-mail or token in any request, withdrawal works across tabs). Use when adding, replacing or auditing web analytics, cookieless tracking, a GDPR\u002FePrivacy consent banner, PostHog privacy settings, or visit notifications, on any stack (plain HTML, Astro, Next.js, React SPAs, Cloudflare, Node).","hashgraph-online\u002Fawesome-codex-plugins","hashgraph-online","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F160371962?v=4","01M6CNB6NRJVNNGN14TR01Q6D3","HOL (Hashgraph Online)","01M6JT1HSKKG9V68KQAQGBA7XP",1575,"MIT","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002FSKILL.md",[18,23,26],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000009","前端开发","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000013","分析与可视化",{"categoryNo":27,"categoryName":28,"taxonomyType":29,"taxonomyTypeDesc":30},"01JBSEEDCAT000000000000033","数据分析师","PROFESSION","适用职业",[5,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51],"stack","section","referral","weekly","deletion","digest","active","show","take","read","slack","astro","Video","ids","cookies","server","notifications","website","analytics","react","2026-11-10 10:12:30",false,"---\nname: sectionsight\ndescription: Sets up privacy-first website analytics that show what visitors actually read (active seconds per page section and whether it was read; scroll depth, CTA and goal clicks, dead and rage clicks, copies, prints, video plays) with no cookies. Adds a consent banner designed around the EDPB Cookie Banner Taskforce report and CNIL guidance, optional PostHog session replay only after consent, a Slack or e-mail line per notable visit, a weekly digest, AI referral and crawler counts, health alerts, enforced retention, and self-service deletion of a visitor's data. Then proves it in a real browser (nothing stored and no ids before consent, no e-mail or token in any request, withdrawal works across tabs). Use when adding, replacing or auditing web analytics, cookieless tracking, a GDPR\u002FePrivacy consent banner, PostHog privacy settings, or visit notifications, on any stack (plain HTML, Astro, Next.js, React SPAs, Cloudflare, Node).\nlicense: MIT\ncompatibility: The reference code needs a server or function that can take a same-origin POST (Node 18+, Workers, Deno, Bun). verify.mjs needs Node 20+, Playwright and Chromium (network access to install them).\nmetadata:\n  author: Jakub Dvořák (kubadvorak.com)\n  version: \"1.0.1\"\n---\n\n# Privacy-first analytics\n\nThis skill wires up website analytics that respect visitors and still tell the\nowner what people read. It bundles the decisions to make, the reference\ncode, and the checks that prove it works, for static sites, server-rendered\nsites and SPAs alike.\n\nThe reference code in `assets\u002F` is tested in a real browser (`scripts\u002Fverify.mjs`).\nCopy what you need into the project, adapt the wiring to its stack, and keep the\ncode's behaviour. Do not rewrite it from memory. Paths below are relative to\nthe directory that contains this SKILL.md.\n\n## Rules that do not bend\n\n1. **Nothing personal leaves the page.** No query string except `utm_*` values\n   that do not look personal, no typed text, no copied text. Path segments\n   that look like an e-mail, phone number, token or id become `[id]`. That\n   rule cannot recognise a name in a slug (`\u002Fu\u002Fjane-doe`), so rule 2 is\n   the real protection for such pages. Every report, event and Slack line\n   follows this.\n2. **Private pages are never measured.** These are URLs that carry a credential\n   or name a person (password reset, invite, payment, `\u002Faccount`, a profile\n   page, the signed-in app). No tracker and no PostHog run there, and no\n   banner shows.\n3. **Nothing is stored in the browser before consent.** The exception is the\n   consent answer itself. Global Privacy Control and Do Not Track mean no.\n4. **The collector never stores an IP address or a full user agent.** Use\n   the CDN's country header and a device class. The collector uses an IP\n   only in memory, for throttling and for a daily hash in model C. If you use\n   the PostHog proxy or `umamiForwarder`: the proxy forwards the full IP to\n   PostHog (turn on \"Discard client IP data\" in the project), and the\n   forwarder sends Umami the IP shortened to its \u002F24 with the full user agent; say so in the\n   privacy text.\n5. **Nothing in Slack or e-mail that identifies a visitor.** A line says\n   where a visit came from and what it did, never who it was.\n6. **Legal choices belong to the owner.** Ask about the consent model, retention\n   and anything else marked *ask*. Never write \"GDPR compliant\" anywhere. Say what\n   was built, and that a lawyer should review the policy text. If nobody can\n   answer (an unattended run), stop after Discover and write the questions\n   into your report. If the owner answered the consent model, storage,\n   retention and private paths up front, build with the recommended\n   defaults for the rest and list each one as unconfirmed.\n7. **Secrets stay secret.** Slack tokens, webhook URLs, PostHog personal API\n   keys and hashing secrets go in environment variables. Never commit them,\n   never print them, never put them in client code. The PostHog *project* token\n   is public by design.\n\n## Workflow\n\n### 1. Discover (read the project, change nothing)\n\n- Stack, rendering (static, server, client-side routing), hosting and CDN.\n  The CDN determines the country header; see `references\u002Fcollector.md`.\n- Where a same-origin POST endpoint can live (route handler, function, server).\n  If the site has none, stop and ask the owner before building anything\n  (`references\u002Frecipes.md`, \"When something does not fit\").\n- What runs today: grep for `gtag`, `googletagmanager`, `posthog`, `plausible`,\n  `umami`, `hotjar`, `clarity`, `fbq`, `_ga`, `document.cookie`, `localStorage`.\n  List what each one sets or sends.\n- Private paths: routes with tokens or ids in the URL, routes whose path\n  names a person (profile or invite slugs), auth and account pages,\n  checkout, admin, the app behind a login.\n- The page structure: `\u003Cmain>`, its `\u003Csection>`s and headings. Note the\n  languages and whether a privacy page exists.\n\n### 2. Decide with the owner\n\nAsk the questions in `references\u002Fdecisions.md`, with its recommended\ndefaults. At minimum:\n\n- **The consent model:**\n  - **A** totals only, usually without a banner (the owner accepts the residual risk);\n  - **B** totals for everyone, per-visit rows (each report stored with a random visit id) and replay only after Accept (needs a banner);\n  - **C** per-visit rows without a banner (more legal exposure; the owner decides).\n- Where the data lives.\n- PostHog replay: yes or no (only with B).\n- Notifications: Slack, e-mail or none.\n- Retention: totals, rows, and how many days Slack visit lines stay\n  (recommend 30, at most 90).\n- Model B: offer \"Delete my visit data\" in the reopened banner, or not.\n- Health alerts: where they go (an ops channel or webhook, never the visits\n  channel), thresholds, quiet hours (`references\u002Falerts.md`).\n- JavaScript errors (`data-errors`) and Core Web Vitals (`data-vitals`):\n  on or off (both off by default; totals only, never ids). Forwarding\n  errors to GlitchTip, Bugsink or Sentry only after consent, if at all\n  (`references\u002Fdecisions.md`, 4d).\n- Whether a known `utm_source` may count as a source in the totals\n  (`campaignTotals`; CNIL's self-assessment profile excludes campaign data),\n  and which `utm_campaign` names, if any, may be counted (`campaigns`).\n  Whether k.js removes `utm_*` from the address bar after reading them\n  (`data-strip-utm`; off by default).\n- The weekly digest (`assets\u002Fdigest.mjs`): yes or no, where it goes, and\n  whether its counts are rounded (default: under 10 shows `\u003C10`, others to\n  tens) or exact (`references\u002Fdecisions.md`, \"Small counts\").\n- Counting the consent banner's answers (`data-consent-counts`, the\n  collector's `consentCounts`): **the owner's decision, off by default**\n  (`references\u002Fdecisions.md`, 4e). It lets the digest say what share of\n  visitors the per-visit rows cover; it is totals only, never ids.\n- AI visibility, each optional (`references\u002Fai-visibility.md`): AI\n  crawler counts on the server (`assets\u002Fcrawlers.mjs`; on Cloudflare its\n  AI Crawl Control is the better source), and Search Console's daily top\n  queries and pages (`assets\u002Fgsc.mjs`; needs a service account added to the\n  property, or Plausible's own integration instead), and whether the\n  digest may list search queries (`gscQueries`; a person's name that is\n  searched often enough can pass Google's anonymisation threshold). Deploy markers from CI\n  (`assets\u002Fci\u002Fdeploy-marker.yml`): yes or no.\n- With `data-errors`: whether alerts and the digest may quote an error's\n  scrubbed message (`showErrorText`, off by default: a name written into a\n  message without quotes is stored and would be posted as written).\n- The time zone the days are counted in.\n- The private paths.\n- **What must never break** (for verify's broken-things checks): the CTA\n  links and where each must point (a sign-up path, a booking or WhatsApp\n  link with the right number), which pages must be indexable by search\n  engines, and which must not be (thank-you pages, internal search). If\n  the owner has no answer, propose the main CTAs found in Discover and\n  the pages in the sitemap. For \"must not be\", name only pages that already\n  carry `noindex` and that robots.txt does not block: a blocked page's\n  `noindex` is never seen (verify fails it), and pages robots.txt already\n  disallows belong in neither list.\n\nWrite the answers into the project (e.g. `ANALYTICS_DECISIONS.md`); they go\ninto the configuration and the privacy text.\n\n### 3. Build\n\nCopy the assets you need from `assets\u002F` into the project, and follow the\nrecipe for the stack in `references\u002Frecipes.md`. Write the private paths as\n**one regex** and use it in all four places: k.js `data-skip`, the\ncollector's `skip`, consent.js `skip`, and posthog-guard's `private`. Paths\nare case-sensitive, so prefer the string form with the `i` flag,\n`\u002F^\\\u002F(account|reset)(\\\u002F|$)\u002Fi`, which all four read. Never a comma list:\n`^\u002Faccount, ^\u002Freset` is rejected; join paths with `|`. A rule that does not\nparse fails closed, differently in each file:\n\n- **k.js** measures nothing and logs a console warning.\n- **consent.js** treats every page as private: the banner shows only when\n  \"Privacy settings\" opens it, so nobody is asked. That is silent in\n  production; on localhost, `127.0.0.1`, `*.localhost` or with `?k-debug`\n  it shows the error at the top of the page.\n- **The collector** throws when it is created. An Express server then fails\n  at boot; a Next.js route that creates it at the top level fails every\n  POST with a 500 (and `next build` may fail); the lazy Cloudflare recipe\n  answers 500 to every report, silently.\n- **posthog-guard** treats every page as private, and `startPostHog` logs an\n  error and never starts PostHog.\n\nSo a broken rule never measures a private page, but it can go unnoticed:\nrun `scripts\u002Fverify.mjs` (step 4) after every change to the rule: \"the\ncollector receives reports\" fails when the collector answers an error, and\n\"the banner is shown\" when the banner never asks.\nThen, in order:\n\n1. **Collector.** Mount `assets\u002Fcollector.mjs` at a same-origin POST path\n   (default `\u002Fapi\u002Fk`). Give it a store, the site's hosts, `skip` (the private\n   paths), `ids` (A `never`, B `consent`, C `always` with a secret) and the\n   time zone. See `references\u002Fcollector.md`.\n2. **Tracker.** Serve `assets\u002Fk.js` from the site, or bundle it, and configure\n   it with the same `ids` and `skip`. Mark the sections with `data-k`, the\n   important links with `data-kcta`, goals with `data-kgoal`, forms with\n   `data-kform`, and accordions, tabs and toggles with `data-kui`. For client-side routing it\n   follows `history` on its own. Add `data-errors=\"on\"` and\n   `data-vitals=\"on\"` only if the owner said yes. See `references\u002Ftracker.md`.\n3. **The 404 page:** put `data-k404` on the `\u003Chtml>` or `\u003Cbody>` of the 404\n   template (`references\u002Ftracker.md`), so broken links show up as\n   `nfin.\u003Cpath>` in the totals. A missing path must answer HTTP 404, not\n   200 with the home page or an app shell: static hosts (Cloudflare Pages,\n   Netlify, Vercel, GitHub Pages) do this when a top-level `404.html`\n   exists and no SPA fallback is configured; Node servers set the status\n   themselves; Next.js does it with `not-found`. An SPA served by its own\n   server answers 404 for paths outside its route list. verify fails a\n   missing path that answers 200.\n4. **Consent (B):** `assets\u002Fconsent.js`, or a port that keeps its behaviour.\n   Add a \"Privacy settings\" control (`data-consent-open`) in the footer of\n   every page, including the 404 page. The banner text must name what Accept\n   turns on and how long data is kept. See `references\u002Fconsent.md`.\n5. **PostHog (B, optional):** `assets\u002Fposthog-guard.mjs`,\n   `assets\u002Fposthog-proxy.mjs` and the project settings in `references\u002Fposthog.md`.\n   It loads only after Accept and stops in the right order on withdrawal.\n6. **Notifications (optional):** `assets\u002Fvisits.mjs` on a schedule. See\n   `references\u002Fnotifications.md`.\n7. **Retention:** `assets\u002Fretention.mjs` once a day for every store without\n   a TTL (files, SQL), and `deleteOldSlackLines` for the Slack lines. Show\n   the owner `retentionStatus`. See `references\u002Fcollector.md`, \"Retention\".\n8. **Health alerts (if wanted):** `assets\u002Fhealth.mjs` every 15–30 minutes,\n   posting to the ops channel. See `references\u002Falerts.md`.\n   **Weekly digest (if wanted):** `assets\u002Fdigest.mjs` once a week, with\n   the owner's rounding choice (`references\u002Fnotifications.md`). **Deploy\n   markers (if wanted):** `assets\u002Fci\u002Fdeploy-marker.yml` and a small\n   `mark-deploy` script that calls `markDeploy`.\n   **AI visibility (the pieces the owner chose):** `countCrawler` from\n   `assets\u002Fcrawlers.mjs` after each response (`waitUntil`, `after()` or a\n   `finish` listener, never awaited before the answer), and `gscDaily` from\n   `assets\u002Fgsc.mjs` once a day. See `references\u002Fai-visibility.md`.\n9. **Deletion:** the collector takes `{ del }` requests; consent.js offers\n   \"Delete my visit data\" unless the owner said no, or the store cannot\n   delete (Analytics Engine): then `deleteButton: false` and k.js\n   `data-delete=\"off\"`.\n   `assets\u002Fdsar.mjs` for requests by e-mail.\n10. **Privacy page:** the checklist in `references\u002Fprivacy-policy.md`, in the\n   site's languages. List only what was actually built, the periods that\n   are actually enforced, and how deletion really works.\n11. **Headers:** `connect-src 'self'` and `script-src 'self'` are enough for\n   k.js. Inline config scripts need a nonce or hash in `script-src`, or an\n   external config file. consent.js's look: on a static host (no\n   per-request nonce), serve `assets\u002Fconsent.css` and set `stylesheet:\n   '\u002Fconsent.css'`, or put its rules in the site's own CSS with `styles:\n   false`; with a per-request nonce, give consent.js's tag the nonce (its\n   `\u003Cstyle>` inherits it). verify fails a CSP violation from these scripts\n   (`references\u002Fconsent.md`). With\n   PostHog heatmaps, allow framing by the PostHog app on public pages only\n   (`references\u002Fposthog.md`).\n\n### 4. Verify (required)\n\nRun `scripts\u002Fverify.mjs` against the running site and fix the code until\nevery check passes. A check never passes because it was weakened. Then do\nthe manual checks in `references\u002Fverify.md` and run the project's own tests.\n\nInstall Playwright in a scratch directory, not in the project: it would\nchange the project's lockfile, and the report lands in the current directory.\n`--resolve example.com` checks the local server under its production name.\n\n```bash\nmkdir -p \u002Ftmp\u002Fsectionsight-verify && cd \u002Ftmp\u002Fsectionsight-verify && npm i playwright && npx playwright install chromium\nnode \u003Cthis skill's directory>\u002Fscripts\u002Fverify.mjs --url http:\u002F\u002Fexample.com:3000 --resolve example.com \\\n  --pages \u002F,\u002Fpricing --private \u002Faccount\u002Fx --mode consent --posthog \u002Fingest\n```\n\nIf verify cannot run (no network, no browser), say so in the report. Never\nreport a check as passed that did not run.\n\nWith `data-consent-counts=\"on\"`, add `--consent-counts`: the banner's\nreports must carry no ids, including the Accept and the withdrawal.\n\nWith `data-errors=\"on\"`, add `--errors`: verify throws an error carrying the\nprobe e-mail and token from an inline script on the first page and checks\nthat the `err` report arrives scrubbed and without ids, before and after\nconsent.\n\nAdd the owner's answers about what must never break: `--expect` for each\nCTA (`'a[data-kcta=signup]@href=^\u002Fsignup$'`), `--indexable` and\n`--noindex` for the pages, `--sitemap`, and `--privacy-page` with\n`--processors` for every recipient the browser does not see (a server that\nforwards to Umami). These checks are read-only. A\ncommand-line `--expect` is checked on every `--pages` page; for a CTA that\nlives on some pages only, use the config file's `expect` entries with their\nown `pages` (`references\u002Fverify.md`).\n\n**Keep it checked (CI).** Write the settings you just used into\n`privacy-check.config.json` (`assets\u002Fci\u002Fprivacy-check.config.example.json`;\nthe URL and `--resolve` stay on the command line), so CI and local runs\nshare them. Then **ask the owner**:\n\n- **On a failed check in a pull request: block the merge, or warn only?**\n  Blocking makes the job a required status check; warning keeps it green\n  and shows the table in the job summary.\n- **Run the check every night against production?** It runs with\n  `--no-write`, so it stores nothing. If yes, **who hears about a failure?**\n  GitHub's own e-mail to whoever last edited the schedule is the default;\n  a Slack webhook step is the alternative.\n\nThen copy `assets\u002Fci\u002Fprivacy-check.yml` to `.github\u002Fworkflows\u002F`, fill in its\n`env` block (production URL; install, build and start commands and the\nlocal URL; `WARN_ONLY`), and drop the nightly job if the owner said no. If\nthe host posts GitHub deployment statuses for previews, tell the owner to\nset the repository variable `PRIVACY_CHECK_PREVIEW=deployment`. Details:\n`references\u002Fverify.md`, \"In CI\". For links across the whole site, also\noffer `assets\u002Fci\u002Flinks.yml` (lychee: internal links on pull requests,\nexternal ones weekly); for uptime and certificate or domain expiry, suggest\na monitor (`references\u002Fdecisions.md`, 4c) rather than building anything,\nand list it in the owner's to-do list. On another CI, run the same command line\n(`--config … --json …`) and keep `privacy-check.json`.\n\n### 5. Report to the owner\n\nWrite a short summary:\n\n- What was built.\n- Which decisions the owner made.\n- The verify output, all of it.\n- What the owner still has to do, for example:\n  - set environment variables;\n  - create the Slack app (and the ops channel for alerts);\n  - schedule the retention and health jobs;\n  - sign the PostHog DPA;\n  - change PostHog project settings;\n  - have the privacy text reviewed;\n  - open the site once with the owner opt-out (`?k=off`) in their own\n    browsers.\n\n## Reference\n\n| File | Use it for |\n|---|---|\n| `references\u002Fdecisions.md` | The questions to ask, the three consent models, defaults |\n| `references\u002Frecipes.md` | How the pieces fit together on plain HTML + Node, Next.js, Astro on Cloudflare |\n| `references\u002Ftracker.md` | What k.js measures, the wire format, config, per-framework setup |\n| `references\u002Fcollector.md` | Validation, storage schema, stores, hosting recipes, reading totals |\n| `references\u002Fconsent.md` | Banner rules, consent.js, porting it to a framework, withdrawal |\n| `references\u002Fposthog.md` | Hardened PostHog: config, before_send, proxy, project settings |\n| `references\u002Fnotifications.md` | Per-visit lines in Slack or e-mail, the weekly digest, deploy markers, a daily-digest recipe |\n| `references\u002Fai-visibility.md` | Visits from AI assistants, AI crawler counts, Search Console's daily top queries |\n| `references\u002Falerts.md` | Health and anomaly alerts: rules, defaults, scheduling, what to ask |\n| `references\u002Fprivacy-policy.md` | What the privacy page must say, per model |\n| `references\u002Fverify.md` | Running verify.mjs, what each check means, the config file, CI, manual checks |\n| `references\u002Fgotchas.md` | Pitfalls and their fixes |\n| `assets\u002F` | k.js, collector.mjs (stores: memory, files, Mongo, SQL, Analytics Engine), visits.mjs, digest.mjs, health.mjs, retention.mjs, dsar.mjs, crawlers.mjs (+ crawlers-list.mjs), gsc.mjs, consent.js, posthog-guard.mjs, posthog-proxy.mjs |\n| `assets\u002Fci\u002F` | GitHub Actions workflows: `privacy-check.yml` (verify on pull requests and nightly), `links.yml` (lychee link check), `deploy-marker.yml` (deploy markers); `privacy-check.config.example.json` |\n| `scripts\u002F` | verify.mjs (the browser check), update-crawlers.mjs (refreshes the crawler list) |\n\n## Mistakes to avoid (details in references\u002Fgotchas.md)\n\n- **A framework rewrite as the PostHog proxy.** It forwards every header,\n  including login cookies. Use the header allowlist proxy.\n- **Calling PostHog `reset()` after `opt_out_capturing()`.** `reset()` clears\n  the opt-out, so withdrawal silently re-enables capture. Opt out last.\n- **Event batching on.** A queued batch is sent after the visitor withdraws.\n  Set `request_batching: false`.\n- **`maskCapturedNetworkRequestFn: () => null`.** It also blanks the page URLs\n  in recordings. Use the selective version in `posthog-guard.mjs`.\n- **Recordings keep links as full URLs.** rrweb stores `\u003Ca href=\"#faq\">` as the\n  whole page URL, including the query. Clean them with the guard's `maskAttributeFn`.\n- **PostHog feature flags left on.** `\u002Fflags` carries the first URL a browser\n  saw, outside `before_send`. Disable flags unless the site uses them.\n- **The entry sent only with the first view.** Consent usually comes after\n  that view, so the per-visit rows lack source, country and device. k.js\n  repeats the entry once (`en: 2`) on the first report with ids.\n- **A path rule only for lower case.** Paths are case-sensitive; write the\n  private-path rule as `\u002F^\\\u002F(account|reset)(\\\u002F|$)\u002Fi` (k.js, consent.js,\n  the collector and posthog-guard all read that form) or redirect other\n  spellings.\n- **Testing in an automated browser without hiding automation.** k.js and\n  PostHog stay silent under `navigator.webdriver`, and collectors drop\n  \"HeadlessChrome\". verify.mjs handles this; hand-written tests must too.\n- **Slack lines that quote app data.** Post only what the visitor did;\n  never mirror message text, form input or data from third-party\n  integrations.\n","# Privacy-first analytics\n\nThis skill wires up website analytics that respect visitors and still tell the\nowner what people read. It bundles the decisions to make, the reference\ncode, and the checks that prove it works, for static sites, server-rendered\nsites and SPAs alike.\n\nThe reference code in `assets\u002F` is tested in a real browser (`scripts\u002Fverify.mjs`).\nCopy what you need into the project, adapt the wiring to its stack, and keep the\ncode's behaviour. Do not rewrite it from memory. Paths below are relative to\nthe directory that contains this SKILL.md.\n\n## Rules that do not bend\n\n1. **Nothing personal leaves the page.** No query string except `utm_*` values\n   that do not look personal, no typed text, no copied text. Path segments\n   that look like an e-mail, phone number, token or id become `[id]`. That\n   rule cannot recognise a name in a slug (`\u002Fu\u002Fjane-doe`), so rule 2 is\n   the real protection for such pages. Every report, event and Slack line\n   follows this.\n2. **Private pages are never measured.** These are URLs that carry a credential\n   or name a person (password reset, invite, payment, `\u002Faccount`, a profile\n   page, the signed-in app). No tracker and no PostHog run there, and no\n   banner shows.\n3. **Nothing is stored in the browser before consent.** The exception is the\n   consent answer itself. Global Privacy Control and Do Not Track mean no.\n4. **The collector never stores an IP address or a full user agent.** Use\n   the CDN's country header and a device class. The collector uses an IP\n   only in memory, for throttling and for a daily hash in model C. If you use\n   the PostHog proxy or `umamiForwarder`: the proxy forwards the full IP to\n   PostHog (turn on \"Discard client IP data\" in the project), and the\n   forwarder sends Umami the IP shortened to its \u002F24 with the full user agent; say so in the\n   privacy text.\n5. **Nothing in Slack or e-mail that identifies a visitor.** A line says\n   where a visit came from and what it did, never who it was.\n6. **Legal choices belong to the owner.** Ask about the consent model, retention\n   and anything else marked *ask*. Never write \"GDPR compliant\" anywhere. Say what\n   was built, and that a lawyer should review the policy text. If nobody can\n   answer (an unattended run), stop after Discover and write the questions\n   into your report. If the owner answered the consent model, storage,\n   retention and private paths up front, build with the recommended\n   defaults for the rest and list each one as unconfirmed.\n7. **Secrets stay secret.** Slack tokens, webhook URLs, PostHog personal API\n   keys and hashing secrets go in environment variables. Never commit them,\n   never print them, never put them in client code. The PostHog *project* token\n   is public by design.\n\n## Workflow\n\n### 1. Discover (read the project, change nothing)\n\n- Stack, rendering (static, server, client-side routing), hosting and CDN.\n  The CDN determines the country header; see `references\u002Fcollector.md`.\n- Where a same-origin POST endpoint can live (route handler, function, server).\n  If the site has none, stop and ask the owner before building anything\n  (`references\u002Frecipes.md`, \"When something does not fit\").\n- What runs today: grep for `gtag`, `googletagmanager`, `posthog`, `plausible`,\n  `umami`, `hotjar`, `clarity`, `fbq`, `_ga`, `document.cookie`, `localStorage`.\n  List what each one sets or sends.\n- Private paths: routes with tokens or ids in the URL, routes whose path\n  names a person (profile or invite slugs), auth and account pages,\n  checkout, admin, the app behind a login.\n- The page structure: `\u003Cmain>`, its `\u003Csection>`s and headings. Note the\n  languages and whether a privacy page exists.\n\n### 2. Decide with the owner\n\nAsk the questions in `references\u002Fdecisions.md`, with its recommended\ndefaults. At minimum:\n\n- **The consent model:**\n  - **A** totals only, usually without a banner (the owner accepts the residual risk);\n  - **B** totals for everyone, per-visit rows (each report stored with a random visit id) and replay only after Accept (needs a banner);\n  - **C** per-visit rows without a banner (more legal exposure; the owner decides).\n- Where the data lives.\n- PostHog replay: yes or no (only with B).\n- Notifications: Slack, e-mail or none.\n- Retention: totals, rows, and how many days Slack visit lines stay\n  (recommend 30, at most 90).\n- Model B: offer \"Delete my visit data\" in the reopened banner, or not.\n- Health alerts: where they go (an ops channel or webhook, never the visits\n  channel), thresholds, quiet hours (`references\u002Falerts.md`).\n- JavaScript errors (`data-errors`) and Core Web Vitals (`data-vitals`):\n  on or off (both off by default; totals only, never ids). Forwarding\n  errors to GlitchTip, Bugsink or Sentry only after consent, if at all\n  (`references\u002Fdecisions.md`, 4d).\n- Whether a known `utm_source` may count as a source in the totals\n  (`campaignTotals`; CNIL's self-assessment profile excludes campaign data),\n  and which `utm_campaign` names, if any, may be counted (`campaigns`).\n  Whether k.js removes `utm_*` from the address bar after reading them\n  (`data-strip-utm`; off by default).\n- The weekly digest (`assets\u002Fdigest.mjs`): yes or no, where it goes, and\n  whether its counts are rounded (default: under 10 shows `\u003C10`, others to\n  tens) or exact (`references\u002Fdecisions.md`, \"Small counts\").\n- Counting the consent banner's answers (`data-consent-counts`, the\n  collector's `consentCounts`): **the owner's decision, off by default**\n  (`references\u002Fdecisions.md`, 4e). It lets the digest say what share of\n  visitors the per-visit rows cover; it is totals only, never ids.\n- AI visibility, each optional (`references\u002Fai-visibility.md`): AI\n  crawler counts on the server (`assets\u002Fcrawlers.mjs`; on Cloudflare its\n  AI Crawl Control is the better source), and Search Console's daily top\n  queries and pages (`assets\u002Fgsc.mjs`; needs a service account added to the\n  property, or Plausible's own integration instead), and whether the\n  digest may list search queries (`gscQueries`; a person's name that is\n  searched often enough can pass Google's anonymisation threshold). Deploy markers from CI\n  (`assets\u002Fci\u002Fdeploy-marker.yml`): yes or no.\n- With `data-errors`: whether alerts and the digest may quote an error's\n  scrubbed message (`showErrorText`, off by default: a name written into a\n  message without quotes is stored and would be posted as written).\n- The time zone the days are counted in.\n- The private paths.\n- **What must never break** (for verify's broken-things checks): the CTA\n  links and where each must point (a sign-up path, a booking or WhatsApp\n  link with the right number), which pages must be indexable by search\n  engines, and which must not be (thank-you pages, internal search). If\n  the owner has no answer, propose the main CTAs found in Discover and\n  the pages in the sitemap. For \"must not be\", name only pages that already\n  carry `noindex` and that robots.txt does not block: a blocked page's\n  `noindex` is never seen (verify fails it), and pages robots.txt already\n  disallows belong in neither list.\n\nWrite the answers into the project (e.g. `ANALYTICS_DECISIONS.md`); they go\ninto the configuration and the privacy text.\n\n### 3. Build\n\nCopy the assets you need from `assets\u002F` into the project, and follow the\nrecipe for the stack in `references\u002Frecipes.md`. Write the private paths as\n**one regex** and use it in all four places: k.js `data-skip`, the\ncollector's `skip`, consent.js `skip`, and posthog-guard's `private`. Paths\nare case-sensitive, so prefer the string form with the `i` flag,\n`\u002F^\\\u002F(account|reset)(\\\u002F|$)\u002Fi`, which all four read. Never a comma list:\n`^\u002Faccount, ^\u002Freset` is rejected; join paths with `|`. A rule that does not\nparse fails closed, differently in each file:\n\n- **k.js** measures nothing and logs a console warning.\n- **consent.js** treats every page as private: the banner shows only when\n  \"Privacy settings\" opens it, so nobody is asked. That is silent in\n  production; on localhost, `127.0.0.1`, `*.localhost` or with `?k-debug`\n  it shows the error at the top of the page.\n- **The collector** throws when it is created. An Express server then fails\n  at boot; a Next.js route that creates it at the top level fails every\n  POST with a 500 (and `next build` may fail); the lazy Cloudflare recipe\n  answers 500 to every report, silently.\n- **posthog-guard** treats every page as private, and `startPostHog` logs an\n  error and never starts PostHog.\n\nSo a broken rule never measures a private page, but it can go unnoticed:\nrun `scripts\u002Fverify.mjs` (step 4) after every change to the rule: \"the\ncollector receives reports\" fails when the collector answers an error, and\n\"the banner is shown\" when the banner never asks.\nThen, in order:\n\n1. **Collector.** Mount `assets\u002Fcollector.mjs` at a same-origin POST path\n   (default `\u002Fapi\u002Fk`). Give it a store, the site's hosts, `skip` (the private\n   paths), `ids` (A `never`, B `consent`, C `always` with a secret) and the\n   time zone. See `references\u002Fcollector.md`.\n2. **Tracker.** Serve `assets\u002Fk.js` from the site, or bundle it, and configure\n   it with the same `ids` and `skip`. Mark the sections with `data-k`, the\n   important links with `data-kcta`, goals with `data-kgoal`, forms with\n   `data-kform`, and accordions, tabs and toggles with `data-kui`. For client-side routing it\n   follows `history` on its own. Add `data-errors=\"on\"` and\n   `data-vitals=\"on\"` only if the owner said yes. See `references\u002Ftracker.md`.\n3. **The 404 page:** put `data-k404` on the `\u003Chtml>` or `\u003Cbody>` of the 404\n   template (`references\u002Ftracker.md`), so broken links show up as\n   `nfin.\u003Cpath>` in the totals. A missing path must answer HTTP 404, not\n   200 with the home page or an app shell: static hosts (Cloudflare Pages,\n   Netlify, Vercel, GitHub Pages) do this when a top-level `404.html`\n   exists and no SPA fallback is configured; Node servers set the status\n   themselves; Next.js does it with `not-found`. An SPA served by its own\n   server answers 404 for paths outside its route list. verify fails a\n   missing path that answers 200.\n4. **Consent (B):** `assets\u002Fconsent.js`, or a port that keeps its behaviour.\n   Add a \"Privacy settings\" control (`data-consent-open`) in the footer of\n   every page, including the 404 page. The banner text must name what Accept\n   turns on and how long data is kept. See `references\u002Fconsent.md`.\n5. **PostHog (B, optional):** `assets\u002Fposthog-guard.mjs`,\n   `assets\u002Fposthog-proxy.mjs` and the project settings in `references\u002Fposthog.md`.\n   It loads only after Accept and stops in the right order on withdrawal.\n6. **Notifications (optional):** `assets\u002Fvisits.mjs` on a schedule. See\n   `references\u002Fnotifications.md`.\n7. **Retention:** `assets\u002Fretention.mjs` once a day for every store without\n   a TTL (files, SQL), and `deleteOldSlackLines` for the Slack lines. Show\n   the owner `retentionStatus`. See `references\u002Fcollector.md`, \"Retention\".\n8. **Health alerts (if wanted):** `assets\u002Fhealth.mjs` every 15–30 minutes,\n   posting to the ops channel. See `references\u002Falerts.md`.\n   **Weekly digest (if wanted):** `assets\u002Fdigest.mjs` once a week, with\n   the owner's rounding choice (`references\u002Fnotifications.md`). **Deploy\n   markers (if wanted):** `assets\u002Fci\u002Fdeploy-marker.yml` and a small\n   `mark-deploy` script that calls `markDeploy`.\n   **AI visibility (the pieces the owner chose):** `countCrawler` from\n   `assets\u002Fcrawlers.mjs` after each response (`waitUntil`, `after()` or a\n   `finish` listener, never awaited before the answer), and `gscDaily` from\n   `assets\u002Fgsc.mjs` once a day. See `references\u002Fai-visibility.md`.\n9. **Deletion:** the collector takes `{ del }` requests; consent.js offers\n   \"Delete my visit data\" unless the owner said no, or the store cannot\n   delete (Analytics Engine): then `deleteButton: false` and k.js\n   `data-delete=\"off\"`.\n   `assets\u002Fdsar.mjs` for requests by e-mail.\n10. **Privacy page:** the checklist in `references\u002Fprivacy-policy.md`, in the\n   site's languages. List only what was actually built, the periods that\n   are actually enforced, and how deletion really works.\n11. **Headers:** `connect-src 'self'` and `script-src 'self'` are enough for\n   k.js. Inline config scripts need a nonce or hash in `script-src`, or an\n   external config file. consent.js's look: on a static host (no\n   per-request nonce), serve `assets\u002Fconsent.css` and set `stylesheet:\n   '\u002Fconsent.css'`, or put its rules in the site's own CSS with `styles:\n   false`; with a per-request nonce, give consent.js's tag the nonce (its\n   `\u003Cstyle>` inherits it). verify fails a CSP violation from these scripts\n   (`references\u002Fconsent.md`). With\n   PostHog heatmaps, allow framing by the PostHog app on public pages only\n   (`references\u002Fposthog.md`).\n\n### 4. Verify (required)\n\nRun `scripts\u002Fverify.mjs` against the running site and fix the code until\nevery check passes. A check never passes because it was weakened. Then do\nthe manual checks in `references\u002Fverify.md` and run the project's own tests.\n\nInstall Playwright in a scratch directory, not in the project: it would\nchange the project's lockfile, and the report lands in the current directory.\n`--resolve example.com` checks the local server under its production name.\n\n```bash\nmkdir -p \u002Ftmp\u002Fsectionsight-verify && cd \u002Ftmp\u002Fsectionsight-verify && npm i playwright && npx playwright install chromium\nnode \u003Cthis skill's directory>\u002Fscripts\u002Fverify.mjs --url http:\u002F\u002Fexample.com:3000 --resolve example.com \\\n  --pages \u002F,\u002Fpricing --private \u002Faccount\u002Fx --mode consent --posthog \u002Fingest\n```\n\nIf verify cannot run (no network, no browser), say so in the report. Never\nreport a check as passed that did not run.\n\nWith `data-consent-counts=\"on\"`, add `--consent-counts`: the banner's\nreports must carry no ids, including the Accept and the withdrawal.\n\nWith `data-errors=\"on\"`, add `--errors`: verify throws an error carrying the\nprobe e-mail and token from an inline script on the first page and checks\nthat the `err` report arrives scrubbed and without ids, before and after\nconsent.\n\nAdd the owner's answers about what must never break: `--expect` for each\nCTA (`'a[data-kcta=signup]@href=^\u002Fsignup$'`), `--indexable` and\n`--noindex` for the pages, `--sitemap`, and `--privacy-page` with\n`--processors` for every recipient the browser does not see (a server that\nforwards to Umami). These checks are read-only. A\ncommand-line `--expect` is checked on every `--pages` page; for a CTA that\nlives on some pages only, use the config file's `expect` entries with their\nown `pages` (`references\u002Fverify.md`).\n\n**Keep it checked (CI).** Write the settings you just used into\n`privacy-check.config.json` (`assets\u002Fci\u002Fprivacy-check.config.example.json`;\nthe URL and `--resolve` stay on the command line), so CI and local runs\nshare them. Then **ask the owner**:\n\n- **On a failed check in a pull request: block the merge, or warn only?**\n  Blocking makes the job a required status check; warning keeps it green\n  and shows the table in the job summary.\n- **Run the check every night against production?** It runs with\n  `--no-write`, so it stores nothing. If yes, **who hears about a failure?**\n  GitHub's own e-mail to whoever last edited the schedule is the default;\n  a Slack webhook step is the alternative.\n\nThen copy `assets\u002Fci\u002Fprivacy-check.yml` to `.github\u002Fworkflows\u002F`, fill in its\n`env` block (production URL; install, build and start commands and the\nlocal URL; `WARN_ONLY`), and drop the nightly job if the owner said no. If\nthe host posts GitHub deployment statuses for previews, tell the owner to\nset the repository variable `PRIVACY_CHECK_PREVIEW=deployment`. Details:\n`references\u002Fverify.md`, \"In CI\". For links across the whole site, also\noffer `assets\u002Fci\u002Flinks.yml` (lychee: internal links on pull requests,\nexternal ones weekly); for uptime and certificate or domain expiry, suggest\na monitor (`references\u002Fdecisions.md`, 4c) rather than building anything,\nand list it in the owner's to-do list. On another CI, run the same command line\n(`--config … --json …`) and keep `privacy-check.json`.\n\n### 5. Report to the owner\n\nWrite a short summary:\n\n- What was built.\n- Which decisions the owner made.\n- The verify output, all of it.\n- What the owner still has to do, for example:\n  - set environment variables;\n  - create the Slack app (and the ops channel for alerts);\n  - schedule the retention and health jobs;\n  - sign the PostHog DPA;\n  - change PostHog project settings;\n  - have the privacy text reviewed;\n  - open the site once with the owner opt-out (`?k=off`) in their own\n    browsers.\n\n## Reference\n\n| File | Use it for |\n|---|---|\n| `references\u002Fdecisions.md` | The questions to ask, the three consent models, defaults |\n| `references\u002Frecipes.md` | How the pieces fit together on plain HTML + Node, Next.js, Astro on Cloudflare |\n| `references\u002Ftracker.md` | What k.js measures, the wire format, config, per-framework setup |\n| `references\u002Fcollector.md` | Validation, storage schema, stores, hosting recipes, reading totals |\n| `references\u002Fconsent.md` | Banner rules, consent.js, porting it to a framework, withdrawal |\n| `references\u002Fposthog.md` | Hardened PostHog: config, before_send, proxy, project settings |\n| `references\u002Fnotifications.md` | Per-visit lines in Slack or e-mail, the weekly digest, deploy markers, a daily-digest recipe |\n| `references\u002Fai-visibility.md` | Visits from AI assistants, AI crawler counts, Search Console's daily top queries |\n| `references\u002Falerts.md` | Health and anomaly alerts: rules, defaults, scheduling, what to ask |\n| `references\u002Fprivacy-policy.md` | What the privacy page must say, per model |\n| `references\u002Fverify.md` | Running verify.mjs, what each check means, the config file, CI, manual checks |\n| `references\u002Fgotchas.md` | Pitfalls and their fixes |\n| `assets\u002F` | k.js, collector.mjs (stores: memory, files, Mongo, SQL, Analytics Engine), visits.mjs, digest.mjs, health.mjs, retention.mjs, dsar.mjs, crawlers.mjs (+ crawlers-list.mjs), gsc.mjs, consent.js, posthog-guard.mjs, posthog-proxy.mjs |\n| `assets\u002Fci\u002F` | GitHub Actions workflows: `privacy-check.yml` (verify on pull requests and nightly), `links.yml` (lychee link check), `deploy-marker.yml` (deploy markers); `privacy-check.config.example.json` |\n| `scripts\u002F` | verify.mjs (the browser check), update-crawlers.mjs (refreshes the crawler list) |\n\n## Mistakes to avoid (details in references\u002Fgotchas.md)\n\n- **A framework rewrite as the PostHog proxy.** It forwards every header,\n  including login cookies. Use the header allowlist proxy.\n- **Calling PostHog `reset()` after `opt_out_capturing()`.** `reset()` clears\n  the opt-out, so withdrawal silently re-enables capture. Opt out last.\n- **Event batching on.** A queued batch is sent after the visitor withdraws.\n  Set `request_batching: false`.\n- **`maskCapturedNetworkRequestFn: () => null`.** It also blanks the page URLs\n  in recordings. Use the selective version in `posthog-guard.mjs`.\n- **Recordings keep links as full URLs.** rrweb stores `\u003Ca href=\"#faq\">` as the\n  whole page URL, including the query. Clean them with the guard's `maskAttributeFn`.\n- **PostHog feature flags left on.** `\u002Fflags` carries the first URL a browser\n  saw, outside `before_send`. Disable flags unless the site uses them.\n- **The entry sent only with the first view.** Consent usually comes after\n  that view, so the per-visit rows lack source, country and device. k.js\n  repeats the entry once (`en: 2`) on the first report with ids.\n- **A path rule only for lower case.** Paths are case-sensitive; write the\n  private-path rule as `\u002F^\\\u002F(account|reset)(\\\u002F|$)\u002Fi` (k.js, consent.js,\n  the collector and posthog-guard all read that form) or redirect other\n  spellings.\n- **Testing in an automated browser without hiding automation.** k.js and\n  PostHog stay silent under `navigator.webdriver`, and collectors drop\n  \"HeadlessChrome\". verify.mjs handles this; hand-written tests must too.\n- **Slack lines that quote app data.** Post only what the visitor did;\n  never mirror message text, form input or data from third-party\n  integrations.\n",{"name":5,"license":15,"metadata":57,"description":7,"compatibility":60},{"author":58,"version":59},"Jakub Dvořák (kubadvorak.com)","1.0.1","The reference code needs a server or function that can take a same-origin POST (Node 18+, Workers, Deno, Bun). verify.mjs needs Node 20+, Playwright and Chromium (network access to install them).","main","2b63f856ce19c2a797929cacabd9e0fc936812fa","plugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight",491,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins","https:\u002F\u002Fhol.org\u002Fplugins\u002Fbest-codex-plugins","https:\u002F\u002Fgithub.com\u002Fhashgraph-online","https:\u002F\u002Fhol.org","Open-source standards and developer tools for interoperable AI agents: discovery, identity, communication, payments, privacy and security.","2026-11-14 14:06:13",{"source":8,"skillName":5,"commands":72,"prompt":82,"zipAvailable":83},[73,76,79],{"tool":74,"command":75},"npx","npx skills add hashgraph-online\u002Fawesome-codex-plugins --skill sectionsight",{"tool":77,"command":78},"bunx","bunx skills add hashgraph-online\u002Fawesome-codex-plugins --skill sectionsight",{"tool":80,"command":81},"pnpm","pnpm dlx skills add hashgraph-online\u002Fawesome-codex-plugins --skill sectionsight","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002FSKILL.md",true,[85,91,95,99,103,109,113,117,121,125,129,133,137,141,145,149,153,157,161,167,171,175,179,183,187,191,195,199,203,207,211,215],{"path":86,"resourceType":87,"resourceTypeDesc":88,"sizeBytes":89,"sourceHtmlUrl":90},"assets\u002Fci\u002Fdeploy-marker.yml","ASSET","资产文件",2253,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fci\u002Fdeploy-marker.yml",{"path":92,"resourceType":87,"resourceTypeDesc":88,"sizeBytes":93,"sourceHtmlUrl":94},"assets\u002Fci\u002Flinks.yml",3262,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fci\u002Flinks.yml",{"path":96,"resourceType":87,"resourceTypeDesc":88,"sizeBytes":97,"sourceHtmlUrl":98},"assets\u002Fci\u002Fprivacy-check.config.example.json",819,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fci\u002Fprivacy-check.config.example.json",{"path":100,"resourceType":87,"resourceTypeDesc":88,"sizeBytes":101,"sourceHtmlUrl":102},"assets\u002Fci\u002Fprivacy-check.yml",8575,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fci\u002Fprivacy-check.yml",{"path":104,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":107,"sourceHtmlUrl":108},"assets\u002Fcollector.mjs","SCRIPT","脚本",89873,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fcollector.mjs",{"path":110,"resourceType":87,"resourceTypeDesc":88,"sizeBytes":111,"sourceHtmlUrl":112},"assets\u002Fconsent.css",1062,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fconsent.css",{"path":114,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":115,"sourceHtmlUrl":116},"assets\u002Fconsent.js",17977,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fconsent.js",{"path":118,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":119,"sourceHtmlUrl":120},"assets\u002Fcrawlers-list.mjs",12689,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fcrawlers-list.mjs",{"path":122,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":123,"sourceHtmlUrl":124},"assets\u002Fcrawlers.mjs",8017,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fcrawlers.mjs",{"path":126,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":127,"sourceHtmlUrl":128},"assets\u002Fdigest.mjs",18555,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fdigest.mjs",{"path":130,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":131,"sourceHtmlUrl":132},"assets\u002Fdsar.mjs",4314,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fdsar.mjs",{"path":134,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":135,"sourceHtmlUrl":136},"assets\u002Fgsc.mjs",9013,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fgsc.mjs",{"path":138,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":139,"sourceHtmlUrl":140},"assets\u002Fhealth.mjs",25949,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fhealth.mjs",{"path":142,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":143,"sourceHtmlUrl":144},"assets\u002Fk.js",44867,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fk.js",{"path":146,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":147,"sourceHtmlUrl":148},"assets\u002Fposthog-guard.mjs",30504,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fposthog-guard.mjs",{"path":150,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":151,"sourceHtmlUrl":152},"assets\u002Fposthog-proxy.mjs",5113,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fposthog-proxy.mjs",{"path":154,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":155,"sourceHtmlUrl":156},"assets\u002Fretention.mjs",6804,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fretention.mjs",{"path":158,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":159,"sourceHtmlUrl":160},"assets\u002Fvisits.mjs",12579,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fassets\u002Fvisits.mjs",{"path":162,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":165,"sourceHtmlUrl":166},"references\u002Fai-visibility.md","REFERENCE","参考文档",11224,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fai-visibility.md",{"path":168,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":169,"sourceHtmlUrl":170},"references\u002Falerts.md",10589,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Falerts.md",{"path":172,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":173,"sourceHtmlUrl":174},"references\u002Fcollector.md",23935,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fcollector.md",{"path":176,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":177,"sourceHtmlUrl":178},"references\u002Fconsent.md",13431,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fconsent.md",{"path":180,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":181,"sourceHtmlUrl":182},"references\u002Fdecisions.md",17484,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fdecisions.md",{"path":184,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":185,"sourceHtmlUrl":186},"references\u002Fgotchas.md",18754,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fgotchas.md",{"path":188,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":189,"sourceHtmlUrl":190},"references\u002Fnotifications.md",15555,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fnotifications.md",{"path":192,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":193,"sourceHtmlUrl":194},"references\u002Fposthog.md",11846,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fposthog.md",{"path":196,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":197,"sourceHtmlUrl":198},"references\u002Fprivacy-policy.md",10831,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fprivacy-policy.md",{"path":200,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":201,"sourceHtmlUrl":202},"references\u002Frecipes.md",14190,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Frecipes.md",{"path":204,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":205,"sourceHtmlUrl":206},"references\u002Ftracker.md",31639,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Ftracker.md",{"path":208,"resourceType":163,"resourceTypeDesc":164,"sizeBytes":209,"sourceHtmlUrl":210},"references\u002Fverify.md",28370,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Freferences\u002Fverify.md",{"path":212,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":213,"sourceHtmlUrl":214},"scripts\u002Fupdate-crawlers.mjs",6713,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fscripts\u002Fupdate-crawlers.mjs",{"path":216,"resourceType":105,"resourceTypeDesc":106,"sizeBytes":217,"sourceHtmlUrl":218},"scripts\u002Fverify.mjs",70558,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FMajny\u002Fsectionsight\u002Fskills\u002Fsectionsight\u002Fscripts\u002Fverify.mjs",[220,229,237,245,253,261],{"skillNo":221,"name":222,"nameZh":6,"description":223,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":224,"sourceHtmlUrl":225,"categories":226,"keywords":227,"publishedAt":228},"01M7DV332YQMGD8WTRJ9DZQXTB","worktable-review-with-annotations","Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-review-with-annotations\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-review-with-annotations\u002FSKILL.md",[],[],"2026-11-15 14:19:17",{"skillNo":230,"name":231,"nameZh":6,"description":232,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":233,"sourceHtmlUrl":234,"categories":235,"keywords":236,"publishedAt":228},"01M7DV2YGHCZVBEB1AHYB6CGHE","worktable-find-and-synthesize-context","Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-find-and-synthesize-context\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-find-and-synthesize-context\u002FSKILL.md",[],[],{"skillNo":238,"name":239,"nameZh":6,"description":240,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":241,"sourceHtmlUrl":242,"categories":243,"keywords":244,"publishedAt":228},"01M7DV2TQN1H11ME2SAF9MXN7V","worktable-create-or-update-docs","Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-update-docs\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-update-docs\u002FSKILL.md",[],[],{"skillNo":246,"name":247,"nameZh":6,"description":248,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":249,"sourceHtmlUrl":250,"categories":251,"keywords":252,"publishedAt":228},"01M7DV2NE452NEA8TXATKSNAJH","worktable-create-or-manage-records","Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-records\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-records\u002FSKILL.md",[],[],{"skillNo":254,"name":255,"nameZh":6,"description":256,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":257,"sourceHtmlUrl":258,"categories":259,"keywords":260,"publishedAt":228},"01M7DV2HENRBR2KSCVM064WH0Y","worktable-create-or-manage-interactive-html","Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-interactive-html\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-create-or-manage-interactive-html\u002FSKILL.md",[],[],{"skillNo":262,"name":263,"nameZh":6,"description":264,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":265,"sourceHtmlUrl":266,"categories":267,"keywords":268,"publishedAt":228},"01M7DV2CADR5R7RQFFYPNAGR9Y","worktable-collaborate-threads","Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-collaborate-threads\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002Fworktable\u002Fworktable-plugin\u002Fskills\u002Fworktable-collaborate-threads\u002FSKILL.md",[],[]]