From 62fc6b5ba5a79f5e713a47ecb83fedb1a441cf31 Mon Sep 17 00:00:00 2001 From: "posthog[bot]" <206114724+posthog[bot]@users.noreply.github.com> Date: Sat, 29 Aug 2026 14:16:01 +0000 Subject: [PATCH 1/2] fix(experiments): repair dead qualitative-feedback link in replay skill The analyzing-experiment-session-replays skill points readers to `references/qualitative-feedback.md`, but the path resolves relative to that skill, which has no references directory. The file lives under the diagnosing-experiment-results skill. Rewrite the reference as a relative link to the actual file so the handoff opens. Generated-By: PostHog Desktop Task-Id: 56e75e09-a7d2-4f81-84de-79cf34a17788 --- .../skills/analyzing-experiment-session-replays/SKILL.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/products/experiments/skills/analyzing-experiment-session-replays/SKILL.md b/products/experiments/skills/analyzing-experiment-session-replays/SKILL.md index e63ad381a2ea..9ec27c6e9ef3 100644 --- a/products/experiments/skills/analyzing-experiment-session-replays/SKILL.md +++ b/products/experiments/skills/analyzing-experiment-session-replays/SKILL.md @@ -162,7 +162,7 @@ worth checking with the people who produced it. Once per conversation at most; d Default to asking every exposed user rather than one variant: a popover shown to only one arm is itself a difference between the arms, and the response event carries the variant anyway, so the split survives. -→ See `references/qualitative-feedback.md` in [[diagnosing-experiment-results]] +→ See [`references/qualitative-feedback.md`](../diagnosing-experiment-results/references/qualitative-feedback.md) in [[diagnosing-experiment-results]] ## Example interaction From 3cc58278a5277f77e8623134e92295ac44c7cd17 Mon Sep 17 00:00:00 2001 From: Marcel Poelker Date: Tue, 1 Sep 2026 14:32:33 -0400 Subject: [PATCH 2/2] fix(experiments): repair the same dead qualitative-feedback link in two sibling skills The two remaining skills that point readers at the survey handoff carried the same unresolvable path as the one already repaired. Both now use the relative link to the file's owning skill, so the reader opens it instead of hitting a dead path. Generated-By: PostHog Desktop Task-Id: 117555df-8e62-47e7-ba6a-7063f1859da5 --- .../experiments/skills/managing-experiment-lifecycle/SKILL.md | 2 +- .../skills/scanning-experiments-with-replay-vision/SKILL.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/products/experiments/skills/managing-experiment-lifecycle/SKILL.md b/products/experiments/skills/managing-experiment-lifecycle/SKILL.md index 83dd4190fd09..5b6dff57efc2 100644 --- a/products/experiments/skills/managing-experiment-lifecycle/SKILL.md +++ b/products/experiments/skills/managing-experiment-lifecycle/SKILL.md @@ -45,7 +45,7 @@ One optional item worth a single mention at launch, when the change is user-faci short survey, shown when users finish the experimented flow (e.g. triggered by the form's submit event), collects qualitative feedback (a rating, an optional comment) alongside the metrics, from day one. Offer it once as setup advice, drop it if declined, and never let it delay the launch. Do **not** raise it at end or ship-variant time — there it reads as a gate -on rolling out. → See `references/qualitative-feedback.md` in [[diagnosing-experiment-results]] +on rolling out. → See [`references/qualitative-feedback.md`](../diagnosing-experiment-results/references/qualitative-feedback.md) in [[diagnosing-experiment-results]] ### Pause (`experiment-pause`) diff --git a/products/experiments/skills/scanning-experiments-with-replay-vision/SKILL.md b/products/experiments/skills/scanning-experiments-with-replay-vision/SKILL.md index 14d01f81149a..156f43bc239a 100644 --- a/products/experiments/skills/scanning-experiments-with-replay-vision/SKILL.md +++ b/products/experiments/skills/scanning-experiments-with-replay-vision/SKILL.md @@ -229,4 +229,4 @@ The check that does settle it is asking the people. A short survey, triggered wh Two constraints carry over from this skill and matter as much here. Don't name the variant in the survey question, for the same reason the scanner prompt doesn't: the answer stops being evidence about the surface. And prefer asking everyone who completes the flow over targeting one variant, because a popover shown to one variant is a difference between the variants that the experiment is still measuring. -→ See `references/qualitative-feedback.md` in [[diagnosing-experiment-results]] +→ See [`references/qualitative-feedback.md`](../diagnosing-experiment-results/references/qualitative-feedback.md) in [[diagnosing-experiment-results]]