[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-posthog-diagnosing-ci-and-merge-bottlenecks":3,"mdc-g5jhzq-key":50,"related-repo-posthog-diagnosing-ci-and-merge-bottlenecks":1197,"related-org-posthog-diagnosing-ci-and-merge-bottlenecks":1298},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":25,"repoUrl":26,"updatedAt":27,"license":28,"forks":29,"topics":30,"repo":45,"sourceUrl":48,"mdContent":49},"diagnosing-ci-and-merge-bottlenecks","diagnose CI and pull request bottlenecks","Diagnoses CI and pull-request pipeline health for a GitHub repo using the engineering analytics MCP tools — pull-requests (PR list with CI status), workflow-health (per-workflow CI trends), and pr-lifecycle (a single PR's timeline). Use when asked whether CI is getting faster or slower, which GitHub Actions workflow is the slow or flaky long-pole, how long PRs take from open to merge, how an author's merge time compares to the cohort, which open PRs have failing or pending CI, or where a specific pull request is stuck. Triggers on \"engineering analytics\", \"is CI getting slower\", \"slow workflow\", \"flaky CI\", \"time to merge\", \"cycle time\", \"PR throughput\", \"failing checks\", \"where is PR \u003Cn> stuck\", \"CI long pole\", \"what's holding up this PR\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"posthog","PostHog","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fposthog.png",[12,16,19,22],{"name":13,"slug":14,"type":15},"GitHub","github","tag",{"name":17,"slug":18,"type":15},"Analytics","analytics",{"name":20,"slug":21,"type":15},"CI\u002FCD","ci-cd",{"name":23,"slug":24,"type":15},"Debugging","debugging",35568,"https:\u002F\u002Fgithub.com\u002FPostHog\u002Fposthog","2026-07-28T05:34:13.113607",null,2977,[31,32,18,33,34,35,36,37,38,39,40,41,42,43,44],"ab-testing","ai-analytics","cdp","data-warehouse","experiments","feature-flags","javascript","product-analytics","python","react","session-replay","surveys","typescript","web-analytics",{"repoUrl":26,"stars":25,"forks":29,"topics":46,"description":47},[31,32,18,33,34,35,36,37,38,39,40,41,42,43,44],"🦔 PostHog is an all-in-one developer platform for building successful products. We offer product analytics, web analytics, session replay, error tracking, feature flags, experimentation, surveys, data warehouse, a CDP, and an AI product assistant to help debug your code, ship features faster, and keep all your usage and customer data in one stack.","https:\u002F\u002Fgithub.com\u002FPostHog\u002Fposthog\u002Ftree\u002FHEAD\u002Fproducts\u002Fengineering_analytics\u002Fskills\u002Fdiagnosing-ci-and-merge-bottlenecks","---\nname: diagnosing-ci-and-merge-bottlenecks\ndescription: >\n  Diagnoses CI and pull-request pipeline health for a GitHub repo using the engineering analytics MCP tools —\n  pull-requests (PR list with CI status), workflow-health (per-workflow CI trends), and pr-lifecycle (a single PR's\n  timeline). Use when asked whether CI is getting faster or slower, which GitHub Actions workflow is the slow or\n  flaky long-pole, how long PRs take from open to merge, how an author's merge time compares to the cohort, which\n  open PRs have failing or pending CI, or where a specific pull request is stuck. Triggers on \"engineering\n  analytics\", \"is CI getting slower\", \"slow workflow\", \"flaky CI\", \"time to merge\", \"cycle time\", \"PR throughput\",\n  \"failing checks\", \"where is PR \u003Cn> stuck\", \"CI long pole\", \"what's holding up this PR\".\n---\n\n# Diagnosing CI and merge bottlenecks\n\nEngineering analytics treats a pull request like product analytics treats a user: a PR moves through a pipeline\n(`opened → CI → review → merged → deployed`) and the job is to find where it slows down. The surface is **named\nMCP tools** — you call them, you don't write SQL. Dogfooded on `PostHog\u002Fposthog`; the same tools serve\nautonomous agents (e.g. PostHog Desktop) reasoning about their own PRs.\n\n## The tools\n\n- **`pull-requests`** — the PR workhorse. Open PRs plus anything merged or closed since `date_from` (default\n  `-30d`), newest first. Each row carries `author` (nested object: `handle`, `display_name`, `is_bot`), `repo`\n  (nested: `owner`, `name`), `state`, `is_draft`, `labels`, `open_to_merge_seconds`, and a `ci` rollup\n  (`runs` \u002F `passing` \u002F `failing` \u002F `pending`) from the head-SHA join. Answers most PR-level questions:\n  which PRs have failing or pending CI, which are stuck open longest, per-author or per-repo triage, and\n  time-to-merge stats (aggregate `open_to_merge_seconds` over the returned merged rows yourself — median and p95,\n  never a mean).\n- **`workflow-health`** — per-workflow CI health over a window (`date_from` \u002F `date_to`, default last 24 hours):\n  `run_count`, `success_rate`, `p50_seconds`, `p95_seconds`, `last_failure_at`. Answers \"is CI getting faster or\n  slower\" and \"which workflow is the slow or flaky long pole\". There is no built-in trend — call it over two\n  adjacent windows and compare. `success_rate` covers completed runs; `p50_seconds` \u002F `p95_seconds` cover\n  successful runs only (cancelled and failed runs end early and would bias the duration trend). Each is `null`\n  when a window has no qualifying runs — guard for null before comparing two windows (a workflow can have runs\n  in one and none in the other). `run_scope=pull_request` scopes to PR-attributed runs, excluding master\u002Fmain\n  (same-repo PRs only — fork runs carry no PR attribution).\n- **`pr-lifecycle`** — a single PR's timeline: a header plus ordered events — opened, then a CI started\u002Ffinished\n  pair **per workflow run** (many on a multi-workflow repo, interleaved by time), then merged\u002Fclosed. Answers\n  \"where is PR N stuck\". `metric_quality` is `partial`.\n- **`engineering-analytics-flaky-tests`** — the active test-health queue from the per-test CI spans, over a\n  window (`date_from` default `-7d`, max 30 days). Evidence is counted per CI run, never per span or run attempt.\n  `classification` is `confirmed_flake` only where the evidence proves nondeterminism\n  (`same_commit_recovery_run_count > 0`: one commit both failed and passed the test, via a \"Re-run failed jobs\"\n  attempt going green or an in-job retry); `quarantined` means a tolerated failure was recorded while masked;\n  `suspected_regression` means only failures were recorded, which is absence of proof, not proof of a real break.\n  A test qualifies on any same-commit recovery, a quarantined failure, any master\u002Fmain failure, or failures on ≥\n  `min_failed_prs` distinct PRs (`failed_pr_count`). Answers \"what is this failing test costing us\" and picks\n  quarantine candidates. **It does not answer \"which tests are flaky\"**: this queue only sees the main Backend pytest\n  and Frontend Jest suites, and recovery proof only arrives when someone re-runs failed jobs (or a pytest test is hand-marked\n  `@pytest.mark.flaky(reruns=N)`). Counts are absolute signal, never rates: passing runs are mostly not\n  emitted, so there is no honest denominator.\n\nThere is no aggregate time-to-merge tool and no \"counts\" tool — derive those from `pull-requests` (the stuck\u002Ffailing\ncounts, the merge-time percentiles).\n\n## Caveats you must carry into every answer\n\nThese are structural limits of today's snapshot data — state them, don't paper over them.\n\n- **`open_to_merge_seconds` is coarse.** It fuses _draft_ time and _ready-for-review_ time into one figure. Report\n  it as \"open to merge\", never \"cycle time\" or \"review time\". Flag it when long-lived drafts inflate a number.\n- **CI status can be stale.** The CI source syncs on a watermark and does not refresh a run that completes after\n  newer runs land (until the `workflow_run` webhook ships). Treat a `pending` count as unsettled, not as a settled\n  failure; lead with status, not a verdict.\n- **CI for a PR is the head-SHA join, nothing else.** The `ci` rollup reflects only the latest commit's runs. There\n  is no other link between a PR and its checks.\n- **No reviews, approvals, per-check\u002Fjob, or deploys yet.** Don't infer review behaviour or DORA metrics from their\n  absence; that data hasn't landed. `pr-lifecycle` is `partial` for the same reason.\n- **Bots and drafts are present in `pull-requests` output, excluded by convention.** Filter out `author.is_bot`\n  (nested under `author`, not a row-level field) and `is_draft` for throughput \u002F merge-time questions; keep them in\n  for bot-impact questions.\n- **`pull-requests` returns a capped page.** At most `limit` rows (newest first); `truncated` is `true` when more\n  match, and there is no repo or limit filter to narrow the call. When `truncated` is `true`, any percentile or\n  count you derive covers only the newest page — not the whole window — so say so and shrink `date_from` until the\n  real set fits under the cap.\n\n## Choosing a tool\n\n| The question                                           | Tool                                | How                                                                                                                                                                                                                                                                                            |\n| ------------------------------------------------------ | ----------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |\n| Is CI getting slower? Which workflow is the long pole? | `workflow-health`                   | Call over two adjacent windows (e.g. `date_from=-14d`, then `date_from=-28d` `date_to=-14d`); compare `p50_seconds` and `p95_seconds` per workflow. Lead with the median but always check p95 separately — they move independently.                                                            |\n| Which open PRs have failing or pending CI?             | `pull-requests`                     | Keep rows where `ci.failing > 0` or `ci.pending > 0`. `pending` means unsettled (or stale) — not a settled failure.                                                                                                                                                                            |\n| Which PRs are stuck open longest?                      | `pull-requests`                     | Keep `state = open`, not `is_draft`, not `author.is_bot`; sort by `created_at` ascending (oldest first).                                                                                                                                                                                       |\n| How long are PRs taking to merge? Per author?          | `pull-requests`                     | Over merged rows (`merged_at` set, not bot, not draft), aggregate `open_to_merge_seconds` — median and p95. Group by `author.handle` for **cohort context, not a ranking** (per-developer surveillance is an explicit non-goal). Trend it by calling with two `date_from` windows.             |\n| Where is PR N stuck?                                   | `pr-lifecycle`                      | Walk the sorted events: `opened → first CI started`, the CI span (first start → last finish; one pair per workflow), `last CI finished → merged`. The largest gap is the bottleneck. A long open→merge with quick CI points at review\u002Fidle time the `partial` data can't itemize yet — say so. |\n| What is a failing test costing us? What to quarantine? | `engineering-analytics-flaky-tests` | Default window is `-7d`; rows are already ranked by blast radius (master failures, then distinct PRs hit). Report counts, never rates. For \"is it flaky\": only `confirmed_flake` rows are proven, and only for tests hand-marked with reruns.                                                  |\n\n## The high-value chain\n\nMirror how a human investigates: aggregate signal → confirm → concrete PR.\n\n```text\nworkflow-health  (find the slow\u002Fflaky long-pole workflow)\n   → pull-requests  (confirm it's dragging merge time; list the affected PRs)\n      → pr-lifecycle  (open a representative stuck PR and show the gap)\n```\n\n\"CI median rose because `e2e-playwright` p95 doubled; that workflow is the long pole on PR #1234, which sat 47m in\nCI before merging.\"\n\n## Output expectations\n\n- Lead with the verdict in one line, then the supporting numbers.\n- Carry the coarse \u002F partial \u002F staleness caveat whenever the distinction matters.\n- For multi-window or multi-workflow comparisons, a short table beats prose. Report median and p95 side by side —\n  never collapse them into one \"average\".\n\n## What NOT to do\n\n- Don't call `open_to_merge_seconds` cycle time or review time — it's coarse open-to-merge.\n- Don't report a CI count as a settled failure when `pending > 0` — it may be unsettled or stale.\n- Don't infer reviews, approvals, per-check counts, or deploys — that data isn't ingested yet.\n- Don't turn per-author buckets into a leaderboard — they're for finding stuck work, not ranking people.\n- Don't reach for these tools to fetch raw PR contents or diffs — they surface pipeline signal, not the PR thread.\n\n## Persisting an answer\n\nThese tools are ad-hoc reads; they cannot be saved as an insight or subscribed to. When the user wants the same\nnumbers as a saved insight, a dashboard tile, or a scheduled email\u002FSlack delivery, switch to the\n`turning-engineering-analytics-into-insights` skill: the underlying warehouse tables\n(`\u003Cprefix>github_pull_requests` \u002F `\u003Cprefix>github_workflow_runs`, prefix from `engineering-analytics-sources`)\nare directly queryable with HogQL, and that skill carries the curated column semantics plus the\ninsight-create \u002F subscriptions-create workflow.\n",{"data":51,"body":52},{"name":4,"description":6},{"type":53,"children":54},"root",[55,63,94,101,499,511,517,522,717,723,1046,1052,1057,1069,1082,1088,1106,1112,1155,1161],{"type":56,"tag":57,"props":58,"children":59},"element","h1",{"id":4},[60],{"type":61,"value":62},"text","Diagnosing CI and merge bottlenecks",{"type":56,"tag":64,"props":65,"children":66},"p",{},[67,69,76,78,84,86,92],{"type":61,"value":68},"Engineering analytics treats a pull request like product analytics treats a user: a PR moves through a pipeline\n(",{"type":56,"tag":70,"props":71,"children":73},"code",{"className":72},[],[74],{"type":61,"value":75},"opened → CI → review → merged → deployed",{"type":61,"value":77},") and the job is to find where it slows down. The surface is ",{"type":56,"tag":79,"props":80,"children":81},"strong",{},[82],{"type":61,"value":83},"named\nMCP tools",{"type":61,"value":85}," — you call them, you don't write SQL. Dogfooded on ",{"type":56,"tag":70,"props":87,"children":89},{"className":88},[],[90],{"type":61,"value":91},"PostHog\u002Fposthog",{"type":61,"value":93},"; the same tools serve\nautonomous agents (e.g. PostHog Desktop) reasoning about their own PRs.",{"type":56,"tag":95,"props":96,"children":98},"h2",{"id":97},"the-tools",[99],{"type":61,"value":100},"The tools",{"type":56,"tag":102,"props":103,"children":104},"ul",{},[105,263,363,400],{"type":56,"tag":106,"props":107,"children":108},"li",{},[109,118,120,126,128,134,136,142,144,150,152,158,159,165,167,173,175,181,182,188,189,195,196,202,203,209,210,216,218,224,226,232,234,240,241,247,248,254,256,261],{"type":56,"tag":79,"props":110,"children":111},{},[112],{"type":56,"tag":70,"props":113,"children":115},{"className":114},[],[116],{"type":61,"value":117},"pull-requests",{"type":61,"value":119}," — the PR workhorse. Open PRs plus anything merged or closed since ",{"type":56,"tag":70,"props":121,"children":123},{"className":122},[],[124],{"type":61,"value":125},"date_from",{"type":61,"value":127}," (default\n",{"type":56,"tag":70,"props":129,"children":131},{"className":130},[],[132],{"type":61,"value":133},"-30d",{"type":61,"value":135},"), newest first. Each row carries ",{"type":56,"tag":70,"props":137,"children":139},{"className":138},[],[140],{"type":61,"value":141},"author",{"type":61,"value":143}," (nested object: ",{"type":56,"tag":70,"props":145,"children":147},{"className":146},[],[148],{"type":61,"value":149},"handle",{"type":61,"value":151},", ",{"type":56,"tag":70,"props":153,"children":155},{"className":154},[],[156],{"type":61,"value":157},"display_name",{"type":61,"value":151},{"type":56,"tag":70,"props":160,"children":162},{"className":161},[],[163],{"type":61,"value":164},"is_bot",{"type":61,"value":166},"), ",{"type":56,"tag":70,"props":168,"children":170},{"className":169},[],[171],{"type":61,"value":172},"repo",{"type":61,"value":174},"\n(nested: ",{"type":56,"tag":70,"props":176,"children":178},{"className":177},[],[179],{"type":61,"value":180},"owner",{"type":61,"value":151},{"type":56,"tag":70,"props":183,"children":185},{"className":184},[],[186],{"type":61,"value":187},"name",{"type":61,"value":166},{"type":56,"tag":70,"props":190,"children":192},{"className":191},[],[193],{"type":61,"value":194},"state",{"type":61,"value":151},{"type":56,"tag":70,"props":197,"children":199},{"className":198},[],[200],{"type":61,"value":201},"is_draft",{"type":61,"value":151},{"type":56,"tag":70,"props":204,"children":206},{"className":205},[],[207],{"type":61,"value":208},"labels",{"type":61,"value":151},{"type":56,"tag":70,"props":211,"children":213},{"className":212},[],[214],{"type":61,"value":215},"open_to_merge_seconds",{"type":61,"value":217},", and a ",{"type":56,"tag":70,"props":219,"children":221},{"className":220},[],[222],{"type":61,"value":223},"ci",{"type":61,"value":225}," rollup\n(",{"type":56,"tag":70,"props":227,"children":229},{"className":228},[],[230],{"type":61,"value":231},"runs",{"type":61,"value":233}," \u002F ",{"type":56,"tag":70,"props":235,"children":237},{"className":236},[],[238],{"type":61,"value":239},"passing",{"type":61,"value":233},{"type":56,"tag":70,"props":242,"children":244},{"className":243},[],[245],{"type":61,"value":246},"failing",{"type":61,"value":233},{"type":56,"tag":70,"props":249,"children":251},{"className":250},[],[252],{"type":61,"value":253},"pending",{"type":61,"value":255},") from the head-SHA join. Answers most PR-level questions:\nwhich PRs have failing or pending CI, which are stuck open longest, per-author or per-repo triage, and\ntime-to-merge stats (aggregate ",{"type":56,"tag":70,"props":257,"children":259},{"className":258},[],[260],{"type":61,"value":215},{"type":61,"value":262}," over the returned merged rows yourself — median and p95,\nnever a mean).",{"type":56,"tag":106,"props":264,"children":265},{},[266,275,277,282,283,289,291,297,298,304,305,311,312,318,319,325,327,332,334,339,340,345,347,353,355,361],{"type":56,"tag":79,"props":267,"children":268},{},[269],{"type":56,"tag":70,"props":270,"children":272},{"className":271},[],[273],{"type":61,"value":274},"workflow-health",{"type":61,"value":276}," — per-workflow CI health over a window (",{"type":56,"tag":70,"props":278,"children":280},{"className":279},[],[281],{"type":61,"value":125},{"type":61,"value":233},{"type":56,"tag":70,"props":284,"children":286},{"className":285},[],[287],{"type":61,"value":288},"date_to",{"type":61,"value":290},", default last 24 hours):\n",{"type":56,"tag":70,"props":292,"children":294},{"className":293},[],[295],{"type":61,"value":296},"run_count",{"type":61,"value":151},{"type":56,"tag":70,"props":299,"children":301},{"className":300},[],[302],{"type":61,"value":303},"success_rate",{"type":61,"value":151},{"type":56,"tag":70,"props":306,"children":308},{"className":307},[],[309],{"type":61,"value":310},"p50_seconds",{"type":61,"value":151},{"type":56,"tag":70,"props":313,"children":315},{"className":314},[],[316],{"type":61,"value":317},"p95_seconds",{"type":61,"value":151},{"type":56,"tag":70,"props":320,"children":322},{"className":321},[],[323],{"type":61,"value":324},"last_failure_at",{"type":61,"value":326},". Answers \"is CI getting faster or\nslower\" and \"which workflow is the slow or flaky long pole\". There is no built-in trend — call it over two\nadjacent windows and compare. ",{"type":56,"tag":70,"props":328,"children":330},{"className":329},[],[331],{"type":61,"value":303},{"type":61,"value":333}," covers completed runs; ",{"type":56,"tag":70,"props":335,"children":337},{"className":336},[],[338],{"type":61,"value":310},{"type":61,"value":233},{"type":56,"tag":70,"props":341,"children":343},{"className":342},[],[344],{"type":61,"value":317},{"type":61,"value":346}," cover\nsuccessful runs only (cancelled and failed runs end early and would bias the duration trend). Each is ",{"type":56,"tag":70,"props":348,"children":350},{"className":349},[],[351],{"type":61,"value":352},"null",{"type":61,"value":354},"\nwhen a window has no qualifying runs — guard for null before comparing two windows (a workflow can have runs\nin one and none in the other). ",{"type":56,"tag":70,"props":356,"children":358},{"className":357},[],[359],{"type":61,"value":360},"run_scope=pull_request",{"type":61,"value":362}," scopes to PR-attributed runs, excluding master\u002Fmain\n(same-repo PRs only — fork runs carry no PR attribution).",{"type":56,"tag":106,"props":364,"children":365},{},[366,375,377,382,384,390,392,398],{"type":56,"tag":79,"props":367,"children":368},{},[369],{"type":56,"tag":70,"props":370,"children":372},{"className":371},[],[373],{"type":61,"value":374},"pr-lifecycle",{"type":61,"value":376}," — a single PR's timeline: a header plus ordered events — opened, then a CI started\u002Ffinished\npair ",{"type":56,"tag":79,"props":378,"children":379},{},[380],{"type":61,"value":381},"per workflow run",{"type":61,"value":383}," (many on a multi-workflow repo, interleaved by time), then merged\u002Fclosed. Answers\n\"where is PR N stuck\". ",{"type":56,"tag":70,"props":385,"children":387},{"className":386},[],[388],{"type":61,"value":389},"metric_quality",{"type":61,"value":391}," is ",{"type":56,"tag":70,"props":393,"children":395},{"className":394},[],[396],{"type":61,"value":397},"partial",{"type":61,"value":399},".",{"type":56,"tag":106,"props":401,"children":402},{},[403,412,414,419,421,427,429,435,436,442,444,450,452,458,460,466,468,474,476,482,484,489,491,497],{"type":56,"tag":79,"props":404,"children":405},{},[406],{"type":56,"tag":70,"props":407,"children":409},{"className":408},[],[410],{"type":61,"value":411},"engineering-analytics-flaky-tests",{"type":61,"value":413}," — the active test-health queue from the per-test CI spans, over a\nwindow (",{"type":56,"tag":70,"props":415,"children":417},{"className":416},[],[418],{"type":61,"value":125},{"type":61,"value":420}," default ",{"type":56,"tag":70,"props":422,"children":424},{"className":423},[],[425],{"type":61,"value":426},"-7d",{"type":61,"value":428},", max 30 days). Evidence is counted per CI run, never per span or run attempt.\n",{"type":56,"tag":70,"props":430,"children":432},{"className":431},[],[433],{"type":61,"value":434},"classification",{"type":61,"value":391},{"type":56,"tag":70,"props":437,"children":439},{"className":438},[],[440],{"type":61,"value":441},"confirmed_flake",{"type":61,"value":443}," only where the evidence proves nondeterminism\n(",{"type":56,"tag":70,"props":445,"children":447},{"className":446},[],[448],{"type":61,"value":449},"same_commit_recovery_run_count > 0",{"type":61,"value":451},": one commit both failed and passed the test, via a \"Re-run failed jobs\"\nattempt going green or an in-job retry); ",{"type":56,"tag":70,"props":453,"children":455},{"className":454},[],[456],{"type":61,"value":457},"quarantined",{"type":61,"value":459}," means a tolerated failure was recorded while masked;\n",{"type":56,"tag":70,"props":461,"children":463},{"className":462},[],[464],{"type":61,"value":465},"suspected_regression",{"type":61,"value":467}," means only failures were recorded, which is absence of proof, not proof of a real break.\nA test qualifies on any same-commit recovery, a quarantined failure, any master\u002Fmain failure, or failures on ≥\n",{"type":56,"tag":70,"props":469,"children":471},{"className":470},[],[472],{"type":61,"value":473},"min_failed_prs",{"type":61,"value":475}," distinct PRs (",{"type":56,"tag":70,"props":477,"children":479},{"className":478},[],[480],{"type":61,"value":481},"failed_pr_count",{"type":61,"value":483},"). Answers \"what is this failing test costing us\" and picks\nquarantine candidates. ",{"type":56,"tag":79,"props":485,"children":486},{},[487],{"type":61,"value":488},"It does not answer \"which tests are flaky\"",{"type":61,"value":490},": this queue only sees the main Backend pytest\nand Frontend Jest suites, and recovery proof only arrives when someone re-runs failed jobs (or a pytest test is hand-marked\n",{"type":56,"tag":70,"props":492,"children":494},{"className":493},[],[495],{"type":61,"value":496},"@pytest.mark.flaky(reruns=N)",{"type":61,"value":498},"). Counts are absolute signal, never rates: passing runs are mostly not\nemitted, so there is no honest denominator.",{"type":56,"tag":64,"props":500,"children":501},{},[502,504,509],{"type":61,"value":503},"There is no aggregate time-to-merge tool and no \"counts\" tool — derive those from ",{"type":56,"tag":70,"props":505,"children":507},{"className":506},[],[508],{"type":61,"value":117},{"type":61,"value":510}," (the stuck\u002Ffailing\ncounts, the merge-time percentiles).",{"type":56,"tag":95,"props":512,"children":514},{"id":513},"caveats-you-must-carry-into-every-answer",[515],{"type":61,"value":516},"Caveats you must carry into every answer",{"type":56,"tag":64,"props":518,"children":519},{},[520],{"type":61,"value":521},"These are structural limits of today's snapshot data — state them, don't paper over them.",{"type":56,"tag":102,"props":523,"children":524},{},[525,555,580,597,620,659],{"type":56,"tag":106,"props":526,"children":527},{},[528,538,540,546,548,553],{"type":56,"tag":79,"props":529,"children":530},{},[531,536],{"type":56,"tag":70,"props":532,"children":534},{"className":533},[],[535],{"type":61,"value":215},{"type":61,"value":537}," is coarse.",{"type":61,"value":539}," It fuses ",{"type":56,"tag":541,"props":542,"children":543},"em",{},[544],{"type":61,"value":545},"draft",{"type":61,"value":547}," time and ",{"type":56,"tag":541,"props":549,"children":550},{},[551],{"type":61,"value":552},"ready-for-review",{"type":61,"value":554}," time into one figure. Report\nit as \"open to merge\", never \"cycle time\" or \"review time\". Flag it when long-lived drafts inflate a number.",{"type":56,"tag":106,"props":556,"children":557},{},[558,563,565,571,573,578],{"type":56,"tag":79,"props":559,"children":560},{},[561],{"type":61,"value":562},"CI status can be stale.",{"type":61,"value":564}," The CI source syncs on a watermark and does not refresh a run that completes after\nnewer runs land (until the ",{"type":56,"tag":70,"props":566,"children":568},{"className":567},[],[569],{"type":61,"value":570},"workflow_run",{"type":61,"value":572}," webhook ships). Treat a ",{"type":56,"tag":70,"props":574,"children":576},{"className":575},[],[577],{"type":61,"value":253},{"type":61,"value":579}," count as unsettled, not as a settled\nfailure; lead with status, not a verdict.",{"type":56,"tag":106,"props":581,"children":582},{},[583,588,590,595],{"type":56,"tag":79,"props":584,"children":585},{},[586],{"type":61,"value":587},"CI for a PR is the head-SHA join, nothing else.",{"type":61,"value":589}," The ",{"type":56,"tag":70,"props":591,"children":593},{"className":592},[],[594],{"type":61,"value":223},{"type":61,"value":596}," rollup reflects only the latest commit's runs. There\nis no other link between a PR and its checks.",{"type":56,"tag":106,"props":598,"children":599},{},[600,605,607,612,613,618],{"type":56,"tag":79,"props":601,"children":602},{},[603],{"type":61,"value":604},"No reviews, approvals, per-check\u002Fjob, or deploys yet.",{"type":61,"value":606}," Don't infer review behaviour or DORA metrics from their\nabsence; that data hasn't landed. ",{"type":56,"tag":70,"props":608,"children":610},{"className":609},[],[611],{"type":61,"value":374},{"type":61,"value":391},{"type":56,"tag":70,"props":614,"children":616},{"className":615},[],[617],{"type":61,"value":397},{"type":61,"value":619}," for the same reason.",{"type":56,"tag":106,"props":621,"children":622},{},[623,635,637,643,645,650,652,657],{"type":56,"tag":79,"props":624,"children":625},{},[626,628,633],{"type":61,"value":627},"Bots and drafts are present in ",{"type":56,"tag":70,"props":629,"children":631},{"className":630},[],[632],{"type":61,"value":117},{"type":61,"value":634}," output, excluded by convention.",{"type":61,"value":636}," Filter out ",{"type":56,"tag":70,"props":638,"children":640},{"className":639},[],[641],{"type":61,"value":642},"author.is_bot",{"type":61,"value":644},"\n(nested under ",{"type":56,"tag":70,"props":646,"children":648},{"className":647},[],[649],{"type":61,"value":141},{"type":61,"value":651},", not a row-level field) and ",{"type":56,"tag":70,"props":653,"children":655},{"className":654},[],[656],{"type":61,"value":201},{"type":61,"value":658}," for throughput \u002F merge-time questions; keep them in\nfor bot-impact questions.",{"type":56,"tag":106,"props":660,"children":661},{},[662,672,674,680,682,688,689,695,697,702,703,708,710,715],{"type":56,"tag":79,"props":663,"children":664},{},[665,670],{"type":56,"tag":70,"props":666,"children":668},{"className":667},[],[669],{"type":61,"value":117},{"type":61,"value":671}," returns a capped page.",{"type":61,"value":673}," At most ",{"type":56,"tag":70,"props":675,"children":677},{"className":676},[],[678],{"type":61,"value":679},"limit",{"type":61,"value":681}," rows (newest first); ",{"type":56,"tag":70,"props":683,"children":685},{"className":684},[],[686],{"type":61,"value":687},"truncated",{"type":61,"value":391},{"type":56,"tag":70,"props":690,"children":692},{"className":691},[],[693],{"type":61,"value":694},"true",{"type":61,"value":696}," when more\nmatch, and there is no repo or limit filter to narrow the call. When ",{"type":56,"tag":70,"props":698,"children":700},{"className":699},[],[701],{"type":61,"value":687},{"type":61,"value":391},{"type":56,"tag":70,"props":704,"children":706},{"className":705},[],[707],{"type":61,"value":694},{"type":61,"value":709},", any percentile or\ncount you derive covers only the newest page — not the whole window — so say so and shrink ",{"type":56,"tag":70,"props":711,"children":713},{"className":712},[],[714],{"type":61,"value":125},{"type":61,"value":716}," until the\nreal set fits under the cap.",{"type":56,"tag":95,"props":718,"children":720},{"id":719},"choosing-a-tool",[721],{"type":61,"value":722},"Choosing a tool",{"type":56,"tag":724,"props":725,"children":726},"table",{},[727,751],{"type":56,"tag":728,"props":729,"children":730},"thead",{},[731],{"type":56,"tag":732,"props":733,"children":734},"tr",{},[735,741,746],{"type":56,"tag":736,"props":737,"children":738},"th",{},[739],{"type":61,"value":740},"The question",{"type":56,"tag":736,"props":742,"children":743},{},[744],{"type":61,"value":745},"Tool",{"type":56,"tag":736,"props":747,"children":748},{},[749],{"type":61,"value":750},"How",{"type":56,"tag":752,"props":753,"children":754},"tbody",{},[755,815,859,909,967,1011],{"type":56,"tag":732,"props":756,"children":757},{},[758,764,772],{"type":56,"tag":759,"props":760,"children":761},"td",{},[762],{"type":61,"value":763},"Is CI getting slower? Which workflow is the long pole?",{"type":56,"tag":759,"props":765,"children":766},{},[767],{"type":56,"tag":70,"props":768,"children":770},{"className":769},[],[771],{"type":61,"value":274},{"type":56,"tag":759,"props":773,"children":774},{},[775,777,783,785,791,793,799,801,806,808,813],{"type":61,"value":776},"Call over two adjacent windows (e.g. ",{"type":56,"tag":70,"props":778,"children":780},{"className":779},[],[781],{"type":61,"value":782},"date_from=-14d",{"type":61,"value":784},", then ",{"type":56,"tag":70,"props":786,"children":788},{"className":787},[],[789],{"type":61,"value":790},"date_from=-28d",{"type":61,"value":792}," ",{"type":56,"tag":70,"props":794,"children":796},{"className":795},[],[797],{"type":61,"value":798},"date_to=-14d",{"type":61,"value":800},"); compare ",{"type":56,"tag":70,"props":802,"children":804},{"className":803},[],[805],{"type":61,"value":310},{"type":61,"value":807}," and ",{"type":56,"tag":70,"props":809,"children":811},{"className":810},[],[812],{"type":61,"value":317},{"type":61,"value":814}," per workflow. Lead with the median but always check p95 separately — they move independently.",{"type":56,"tag":732,"props":816,"children":817},{},[818,823,831],{"type":56,"tag":759,"props":819,"children":820},{},[821],{"type":61,"value":822},"Which open PRs have failing or pending CI?",{"type":56,"tag":759,"props":824,"children":825},{},[826],{"type":56,"tag":70,"props":827,"children":829},{"className":828},[],[830],{"type":61,"value":117},{"type":56,"tag":759,"props":832,"children":833},{},[834,836,842,844,850,852,857],{"type":61,"value":835},"Keep rows where ",{"type":56,"tag":70,"props":837,"children":839},{"className":838},[],[840],{"type":61,"value":841},"ci.failing > 0",{"type":61,"value":843}," or ",{"type":56,"tag":70,"props":845,"children":847},{"className":846},[],[848],{"type":61,"value":849},"ci.pending > 0",{"type":61,"value":851},". ",{"type":56,"tag":70,"props":853,"children":855},{"className":854},[],[856],{"type":61,"value":253},{"type":61,"value":858}," means unsettled (or stale) — not a settled failure.",{"type":56,"tag":732,"props":860,"children":861},{},[862,867,875],{"type":56,"tag":759,"props":863,"children":864},{},[865],{"type":61,"value":866},"Which PRs are stuck open longest?",{"type":56,"tag":759,"props":868,"children":869},{},[870],{"type":56,"tag":70,"props":871,"children":873},{"className":872},[],[874],{"type":61,"value":117},{"type":56,"tag":759,"props":876,"children":877},{},[878,880,886,888,893,894,899,901,907],{"type":61,"value":879},"Keep ",{"type":56,"tag":70,"props":881,"children":883},{"className":882},[],[884],{"type":61,"value":885},"state = open",{"type":61,"value":887},", not ",{"type":56,"tag":70,"props":889,"children":891},{"className":890},[],[892],{"type":61,"value":201},{"type":61,"value":887},{"type":56,"tag":70,"props":895,"children":897},{"className":896},[],[898],{"type":61,"value":642},{"type":61,"value":900},"; sort by ",{"type":56,"tag":70,"props":902,"children":904},{"className":903},[],[905],{"type":61,"value":906},"created_at",{"type":61,"value":908}," ascending (oldest first).",{"type":56,"tag":732,"props":910,"children":911},{},[912,917,925],{"type":56,"tag":759,"props":913,"children":914},{},[915],{"type":61,"value":916},"How long are PRs taking to merge? Per author?",{"type":56,"tag":759,"props":918,"children":919},{},[920],{"type":56,"tag":70,"props":921,"children":923},{"className":922},[],[924],{"type":61,"value":117},{"type":56,"tag":759,"props":926,"children":927},{},[928,930,936,938,943,945,951,953,958,960,965],{"type":61,"value":929},"Over merged rows (",{"type":56,"tag":70,"props":931,"children":933},{"className":932},[],[934],{"type":61,"value":935},"merged_at",{"type":61,"value":937}," set, not bot, not draft), aggregate ",{"type":56,"tag":70,"props":939,"children":941},{"className":940},[],[942],{"type":61,"value":215},{"type":61,"value":944}," — median and p95. Group by ",{"type":56,"tag":70,"props":946,"children":948},{"className":947},[],[949],{"type":61,"value":950},"author.handle",{"type":61,"value":952}," for ",{"type":56,"tag":79,"props":954,"children":955},{},[956],{"type":61,"value":957},"cohort context, not a ranking",{"type":61,"value":959}," (per-developer surveillance is an explicit non-goal). Trend it by calling with two ",{"type":56,"tag":70,"props":961,"children":963},{"className":962},[],[964],{"type":61,"value":125},{"type":61,"value":966}," windows.",{"type":56,"tag":732,"props":968,"children":969},{},[970,975,983],{"type":56,"tag":759,"props":971,"children":972},{},[973],{"type":61,"value":974},"Where is PR N stuck?",{"type":56,"tag":759,"props":976,"children":977},{},[978],{"type":56,"tag":70,"props":979,"children":981},{"className":980},[],[982],{"type":61,"value":374},{"type":56,"tag":759,"props":984,"children":985},{},[986,988,994,996,1002,1004,1009],{"type":61,"value":987},"Walk the sorted events: ",{"type":56,"tag":70,"props":989,"children":991},{"className":990},[],[992],{"type":61,"value":993},"opened → first CI started",{"type":61,"value":995},", the CI span (first start → last finish; one pair per workflow), ",{"type":56,"tag":70,"props":997,"children":999},{"className":998},[],[1000],{"type":61,"value":1001},"last CI finished → merged",{"type":61,"value":1003},". The largest gap is the bottleneck. A long open→merge with quick CI points at review\u002Fidle time the ",{"type":56,"tag":70,"props":1005,"children":1007},{"className":1006},[],[1008],{"type":61,"value":397},{"type":61,"value":1010}," data can't itemize yet — say so.",{"type":56,"tag":732,"props":1012,"children":1013},{},[1014,1019,1027],{"type":56,"tag":759,"props":1015,"children":1016},{},[1017],{"type":61,"value":1018},"What is a failing test costing us? What to quarantine?",{"type":56,"tag":759,"props":1020,"children":1021},{},[1022],{"type":56,"tag":70,"props":1023,"children":1025},{"className":1024},[],[1026],{"type":61,"value":411},{"type":56,"tag":759,"props":1028,"children":1029},{},[1030,1032,1037,1039,1044],{"type":61,"value":1031},"Default window is ",{"type":56,"tag":70,"props":1033,"children":1035},{"className":1034},[],[1036],{"type":61,"value":426},{"type":61,"value":1038},"; rows are already ranked by blast radius (master failures, then distinct PRs hit). Report counts, never rates. For \"is it flaky\": only ",{"type":56,"tag":70,"props":1040,"children":1042},{"className":1041},[],[1043],{"type":61,"value":441},{"type":61,"value":1045}," rows are proven, and only for tests hand-marked with reruns.",{"type":56,"tag":95,"props":1047,"children":1049},{"id":1048},"the-high-value-chain",[1050],{"type":61,"value":1051},"The high-value chain",{"type":56,"tag":64,"props":1053,"children":1054},{},[1055],{"type":61,"value":1056},"Mirror how a human investigates: aggregate signal → confirm → concrete PR.",{"type":56,"tag":1058,"props":1059,"children":1064},"pre",{"className":1060,"code":1062,"language":61,"meta":1063},[1061],"language-text","workflow-health  (find the slow\u002Fflaky long-pole workflow)\n   → pull-requests  (confirm it's dragging merge time; list the affected PRs)\n      → pr-lifecycle  (open a representative stuck PR and show the gap)\n","",[1065],{"type":56,"tag":70,"props":1066,"children":1067},{"__ignoreMap":1063},[1068],{"type":61,"value":1062},{"type":56,"tag":64,"props":1070,"children":1071},{},[1072,1074,1080],{"type":61,"value":1073},"\"CI median rose because ",{"type":56,"tag":70,"props":1075,"children":1077},{"className":1076},[],[1078],{"type":61,"value":1079},"e2e-playwright",{"type":61,"value":1081}," p95 doubled; that workflow is the long pole on PR #1234, which sat 47m in\nCI before merging.\"",{"type":56,"tag":95,"props":1083,"children":1085},{"id":1084},"output-expectations",[1086],{"type":61,"value":1087},"Output expectations",{"type":56,"tag":102,"props":1089,"children":1090},{},[1091,1096,1101],{"type":56,"tag":106,"props":1092,"children":1093},{},[1094],{"type":61,"value":1095},"Lead with the verdict in one line, then the supporting numbers.",{"type":56,"tag":106,"props":1097,"children":1098},{},[1099],{"type":61,"value":1100},"Carry the coarse \u002F partial \u002F staleness caveat whenever the distinction matters.",{"type":56,"tag":106,"props":1102,"children":1103},{},[1104],{"type":61,"value":1105},"For multi-window or multi-workflow comparisons, a short table beats prose. Report median and p95 side by side —\nnever collapse them into one \"average\".",{"type":56,"tag":95,"props":1107,"children":1109},{"id":1108},"what-not-to-do",[1110],{"type":61,"value":1111},"What NOT to do",{"type":56,"tag":102,"props":1113,"children":1114},{},[1115,1127,1140,1145,1150],{"type":56,"tag":106,"props":1116,"children":1117},{},[1118,1120,1125],{"type":61,"value":1119},"Don't call ",{"type":56,"tag":70,"props":1121,"children":1123},{"className":1122},[],[1124],{"type":61,"value":215},{"type":61,"value":1126}," cycle time or review time — it's coarse open-to-merge.",{"type":56,"tag":106,"props":1128,"children":1129},{},[1130,1132,1138],{"type":61,"value":1131},"Don't report a CI count as a settled failure when ",{"type":56,"tag":70,"props":1133,"children":1135},{"className":1134},[],[1136],{"type":61,"value":1137},"pending > 0",{"type":61,"value":1139}," — it may be unsettled or stale.",{"type":56,"tag":106,"props":1141,"children":1142},{},[1143],{"type":61,"value":1144},"Don't infer reviews, approvals, per-check counts, or deploys — that data isn't ingested yet.",{"type":56,"tag":106,"props":1146,"children":1147},{},[1148],{"type":61,"value":1149},"Don't turn per-author buckets into a leaderboard — they're for finding stuck work, not ranking people.",{"type":56,"tag":106,"props":1151,"children":1152},{},[1153],{"type":61,"value":1154},"Don't reach for these tools to fetch raw PR contents or diffs — they surface pipeline signal, not the PR thread.",{"type":56,"tag":95,"props":1156,"children":1158},{"id":1157},"persisting-an-answer",[1159],{"type":61,"value":1160},"Persisting an answer",{"type":56,"tag":64,"props":1162,"children":1163},{},[1164,1166,1172,1174,1180,1181,1187,1189,1195],{"type":61,"value":1165},"These tools are ad-hoc reads; they cannot be saved as an insight or subscribed to. When the user wants the same\nnumbers as a saved insight, a dashboard tile, or a scheduled email\u002FSlack delivery, switch to the\n",{"type":56,"tag":70,"props":1167,"children":1169},{"className":1168},[],[1170],{"type":61,"value":1171},"turning-engineering-analytics-into-insights",{"type":61,"value":1173}," skill: the underlying warehouse tables\n(",{"type":56,"tag":70,"props":1175,"children":1177},{"className":1176},[],[1178],{"type":61,"value":1179},"\u003Cprefix>github_pull_requests",{"type":61,"value":233},{"type":56,"tag":70,"props":1182,"children":1184},{"className":1183},[],[1185],{"type":61,"value":1186},"\u003Cprefix>github_workflow_runs",{"type":61,"value":1188},", prefix from ",{"type":56,"tag":70,"props":1190,"children":1192},{"className":1191},[],[1193],{"type":61,"value":1194},"engineering-analytics-sources",{"type":61,"value":1196},")\nare directly queryable with HogQL, and that skill carries the curated column semantics plus the\ninsight-create \u002F subscriptions-create workflow.",{"items":1198,"total":1297},[1199,1214,1226,1238,1251,1264,1280],{"slug":1200,"name":1200,"fn":1201,"description":1202,"org":1203,"tags":1204,"stars":25,"repoUrl":26,"updatedAt":1213},"analyzing-expensive-users","analyze expensive users in AI observability","Analyze the most expensive users in AI observability and explain why they cost so much. Use when the user asks about top spenders, expensive users, per-user LLM cost, user-level cost drivers, or patterns behind high AI observability spend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1205,1206,1209,1212],{"name":17,"slug":18,"type":15},{"name":1207,"slug":1208,"type":15},"Cost Optimization","cost-optimization",{"name":1210,"slug":1211,"type":15},"Observability","observability",{"name":9,"slug":8,"type":15},"2026-07-28T05:34:11.117757",{"slug":1215,"name":1215,"fn":1216,"description":1217,"org":1218,"tags":1219,"stars":25,"repoUrl":26,"updatedAt":1225},"auditing-endpoints","audit PostHog project endpoints","Audit every endpoint in a PostHog project for staleness, failed materialisations, and unused materialised versions. Use when the user asks \"what endpoints can I clean up?\", \"are any of my endpoints broken?\", \"which materialised versions are still being called?\", or wants a one-shot cleanup pass over the Endpoints product. Produces a prioritised report grouped by issue type, with recommended actions but does not modify anything without explicit confirmation.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1220,1221,1224],{"name":17,"slug":18,"type":15},{"name":1222,"slug":1223,"type":15},"Audit","audit",{"name":9,"slug":8,"type":15},"2026-06-08T08:08:33.693989",{"slug":1227,"name":1227,"fn":1228,"description":1229,"org":1230,"tags":1231,"stars":25,"repoUrl":26,"updatedAt":1237},"auditing-warehouse-source-health","audit PostHog data warehouse source health","Audit the health of a PostHog project's data warehouse sources and syncs — find every broken or degraded source connection, sync schema, and webhook channel. Use when the user asks \"why are my imports failing?\", \"what's broken with my sources?\", \"why is my warehouse data stale?\", or wants a one-shot triage of source\u002Fsync health before deciding where to dig in. Produces a prioritized report grouped by severity, with recommended next steps. For materialized-view health use `auditing-warehouse-view-health`; for a single failing sync use `diagnosing-failed-warehouse-syncs`.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1232,1233,1235,1236],{"name":1222,"slug":1223,"type":15},{"name":1234,"slug":34,"type":15},"Data Warehouse",{"name":1210,"slug":1211,"type":15},{"name":9,"slug":8,"type":15},"2026-06-18T08:22:57.67984",{"slug":1239,"name":1239,"fn":1240,"description":1241,"org":1242,"tags":1243,"stars":25,"repoUrl":26,"updatedAt":1250},"auditing-warehouse-view-health","audit PostHog materialized view health","Audit the health of a PostHog project's materialized views (saved queries) — find every failed materialization and flag unused or stale materialized views that cost storage and compute. Use when the user asks \"which of my views are broken?\", \"why is this materialized view failing?\", \"are any of my views wasting compute?\", or wants a one-shot triage of view health. For source\u002Fsync health use `auditing-warehouse-source-health`.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1244,1245,1246,1249],{"name":1222,"slug":1223,"type":15},{"name":1234,"slug":34,"type":15},{"name":1247,"slug":1248,"type":15},"Performance","performance",{"name":9,"slug":8,"type":15},"2026-06-18T08:25:10.936787",{"slug":1252,"name":1252,"fn":1253,"description":1254,"org":1255,"tags":1256,"stars":25,"repoUrl":26,"updatedAt":1263},"authoring-error-tracking-alerts","author PostHog error tracking alerts","Author error tracking alerts that fire when an issue is created, reopened, or starts spiking. Use when the user asks to set up error notifications, route exceptions to Slack\u002Fwebhook\u002FLinear, or evaluate which error events are worth alerting on. Covers trigger-event selection, integration choice, dedup against existing alerts, and shipping with the canonical message body shape.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1257,1260,1261,1262],{"name":1258,"slug":1259,"type":15},"Alerting","alerting",{"name":23,"slug":24,"type":15},{"name":1210,"slug":1211,"type":15},{"name":9,"slug":8,"type":15},"2026-06-18T08:24:40.318583",{"slug":1265,"name":1265,"fn":1266,"description":1267,"org":1268,"tags":1269,"stars":25,"repoUrl":26,"updatedAt":1279},"authoring-log-alerts","author log alerts in PostHog","Author useful, low-noise log alerts on services in a PostHog project. Use when the user asks to set up alerts for their logs, suggest alerts they should add, or evaluate whether a service is worth monitoring. Covers service triage, baseline characterisation, threshold drafting, back-testing via simulate, and shipping with a notification destination.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1270,1271,1274,1275,1278],{"name":17,"slug":18,"type":15},{"name":1272,"slug":1273,"type":15},"Monitoring","monitoring",{"name":1210,"slug":1211,"type":15},{"name":1276,"slug":1277,"type":15},"Operations","operations",{"name":9,"slug":8,"type":15},"2026-07-18T05:10:54.430898",{"slug":1281,"name":1281,"fn":1282,"description":1283,"org":1284,"tags":1285,"stars":25,"repoUrl":26,"updatedAt":1296},"building-workflows","build and edit PostHog workflows","Build, edit, test, enable, and monitor PostHog workflows over MCP. Author the action\u002Fedge graph so it runs and opens cleanly in the visual editor, then change drafts surgically with patch operations. Use when asked to build, set up, automate, change, fix, or debug a workflow, campaign, broadcast, drip sequence, or event-triggered automation in the workflows product.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1286,1289,1292,1293],{"name":1287,"slug":1288,"type":15},"Automation","automation",{"name":1290,"slug":1291,"type":15},"MCP","mcp",{"name":9,"slug":8,"type":15},{"name":1294,"slug":1295,"type":15},"Workflow Automation","workflow-automation","2026-07-28T05:34:12.167015",61,{"items":1299,"total":1416},[1300,1307,1313,1320,1327,1334,1342,1349,1363,1378,1388,1406],{"slug":1200,"name":1200,"fn":1201,"description":1202,"org":1301,"tags":1302,"stars":25,"repoUrl":26,"updatedAt":1213},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1303,1304,1305,1306],{"name":17,"slug":18,"type":15},{"name":1207,"slug":1208,"type":15},{"name":1210,"slug":1211,"type":15},{"name":9,"slug":8,"type":15},{"slug":1215,"name":1215,"fn":1216,"description":1217,"org":1308,"tags":1309,"stars":25,"repoUrl":26,"updatedAt":1225},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1310,1311,1312],{"name":17,"slug":18,"type":15},{"name":1222,"slug":1223,"type":15},{"name":9,"slug":8,"type":15},{"slug":1227,"name":1227,"fn":1228,"description":1229,"org":1314,"tags":1315,"stars":25,"repoUrl":26,"updatedAt":1237},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1316,1317,1318,1319],{"name":1222,"slug":1223,"type":15},{"name":1234,"slug":34,"type":15},{"name":1210,"slug":1211,"type":15},{"name":9,"slug":8,"type":15},{"slug":1239,"name":1239,"fn":1240,"description":1241,"org":1321,"tags":1322,"stars":25,"repoUrl":26,"updatedAt":1250},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1323,1324,1325,1326],{"name":1222,"slug":1223,"type":15},{"name":1234,"slug":34,"type":15},{"name":1247,"slug":1248,"type":15},{"name":9,"slug":8,"type":15},{"slug":1252,"name":1252,"fn":1253,"description":1254,"org":1328,"tags":1329,"stars":25,"repoUrl":26,"updatedAt":1263},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1330,1331,1332,1333],{"name":1258,"slug":1259,"type":15},{"name":23,"slug":24,"type":15},{"name":1210,"slug":1211,"type":15},{"name":9,"slug":8,"type":15},{"slug":1265,"name":1265,"fn":1266,"description":1267,"org":1335,"tags":1336,"stars":25,"repoUrl":26,"updatedAt":1279},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1337,1338,1339,1340,1341],{"name":17,"slug":18,"type":15},{"name":1272,"slug":1273,"type":15},{"name":1210,"slug":1211,"type":15},{"name":1276,"slug":1277,"type":15},{"name":9,"slug":8,"type":15},{"slug":1281,"name":1281,"fn":1282,"description":1283,"org":1343,"tags":1344,"stars":25,"repoUrl":26,"updatedAt":1296},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1345,1346,1347,1348],{"name":1287,"slug":1288,"type":15},{"name":1290,"slug":1291,"type":15},{"name":9,"slug":8,"type":15},{"name":1294,"slug":1295,"type":15},{"slug":1350,"name":1350,"fn":1351,"description":1352,"org":1353,"tags":1354,"stars":25,"repoUrl":26,"updatedAt":1362},"check-posthog-loading","inspect PostHog SDK loading across URLs","Inspect how the PostHog JavaScript SDK is loaded across a list of URLs. Use to confirm consistent installation across pages, find pages missing the snippet, detect mismatched API keys or hosts between pages, and verify the load method (head snippet vs deferred vs array.js).\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1355,1356,1357,1360,1361],{"name":17,"slug":18,"type":15},{"name":23,"slug":24,"type":15},{"name":1358,"slug":1359,"type":15},"Frontend","frontend",{"name":1210,"slug":1211,"type":15},{"name":9,"slug":8,"type":15},"2026-05-07T05:56:19.828048",{"slug":1364,"name":1364,"fn":1365,"description":1366,"org":1367,"tags":1368,"stars":25,"repoUrl":26,"updatedAt":1377},"consuming-endpoints-from-client-code","integrate PostHog endpoints into client applications","Wire a PostHog endpoint into a client app or SDK. Covers fetching the OpenAPI spec, generating a typed client with openapi-generator or @hey-api\u002Fopenapi-ts, sending the right auth header, shaping the variables payload (HogQL code_name vs insight breakdown property), handling rate-limit and materialised-endpoint error responses. Use when the user says \"how do I call my endpoint\", \"generate a client for this\", or \"what auth header do I use\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1369,1372,1373,1374],{"name":1370,"slug":1371,"type":15},"API Development","api-development",{"name":1358,"slug":1359,"type":15},{"name":9,"slug":8,"type":15},{"name":1375,"slug":1376,"type":15},"SDK","sdk","2026-06-08T08:08:34.929454",{"slug":1379,"name":1379,"fn":1380,"description":1381,"org":1382,"tags":1383,"stars":25,"repoUrl":26,"updatedAt":1387},"copying-endpoints-across-projects","copy PostHog endpoints across projects","Copy a PostHog endpoint (a saved HogQL\u002Finsight query exposed as an API route) to another project in the same organization, or duplicate it under a new name in the same project. Use when the user wants to duplicate an endpoint, promote an endpoint from staging to production, replicate an endpoint's query\u002Fvariables\u002Ffreshness config in another workspace, or clone an endpoint to iterate on it. Unlike feature flags and experiments, endpoints have NO native cross-project copy tool — this skill covers the read-then-recreate flow (endpoint-get then endpoint-create), the active-project switching it requires, name-collision checks, and the safe defaults (land unmaterialised in the target, verify with endpoint-run). Does not cover editing endpoint versions (see managing-endpoint-versions) or authoring a brand-new endpoint from scratch (see creating-an-endpoint).\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1384,1385,1386],{"name":1370,"slug":1371,"type":15},{"name":1276,"slug":1277,"type":15},{"name":9,"slug":8,"type":15},"2026-07-15T05:29:58.442727",{"slug":1389,"name":1389,"fn":1390,"description":1391,"org":1392,"tags":1393,"stars":25,"repoUrl":26,"updatedAt":1405},"creating-ai-subscription","schedule recurring AI-generated PostHog reports","Create a recurring AI-generated PostHog report — schedule a free-text prompt to run on a cron, with the LLM-synthesized markdown delivered to email or Slack on each tick. Use when the user wants a recurring AI summary of X on any cadence (daily, weekly, monthly, yearly) rather than a one-off report. (To attach an AI summary to an existing insight\u002Fdashboard subscription instead of a free-text prompt, see `managing-subscriptions` and its `summary_enabled` option.)\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1394,1395,1398,1399,1402],{"name":1287,"slug":1288,"type":15},{"name":1396,"slug":1397,"type":15},"Email","email",{"name":9,"slug":8,"type":15},{"name":1400,"slug":1401,"type":15},"Reporting","reporting",{"name":1403,"slug":1404,"type":15},"Slack","slack","2026-06-09T07:32:27.935712",{"slug":1407,"name":1407,"fn":1408,"description":1409,"org":1410,"tags":1411,"stars":25,"repoUrl":26,"updatedAt":1415},"creating-an-endpoint","create PostHog API endpoints","Create a PostHog endpoint with the right shape on the first try — covers query kind choice, name conventions, what to expose as variables (HogQL code_name vs insight breakdown), data_freshness_seconds, and whether to materialise on day one. Use when the user says \"create an endpoint\", \"expose this query as an API\", \"turn this insight into an endpoint\", or asks for help structuring a new endpoint. Steers away from common mistakes: materialising a query with cohort breakdowns or compare mode, inline-only variables on a materialised endpoint, unbounded date ranges, ambiguous names.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1412,1413,1414],{"name":17,"slug":18,"type":15},{"name":1370,"slug":1371,"type":15},{"name":9,"slug":8,"type":15},"2026-06-08T08:08:29.624498",231]