diff --git a/.opencode/skills/app-e2e-tests/SKILL.md b/.opencode/skills/app-e2e-tests/SKILL.md
index ac57961b9f6..d8afe92c366 100644
--- a/.opencode/skills/app-e2e-tests/SKILL.md
+++ b/.opencode/skills/app-e2e-tests/SKILL.md
@@ -122,9 +122,9 @@ Run from `packages/app`:
 
 ## When CI flakes
 
-- **Find it.** e2e runs on pull requests that affect `@opencode/app` or `@opencode/stories`, including their workspace dependencies, and on every manual dispatch (`.github/workflows/check-test.yml:50-62`). It never runs on dev (`:28`).
+- **Find it.** e2e runs on every push to dev and beta, on nightly and manual runs, and as the gate `release.yml` calls before a stable release. Pull requests never run it. On a push the `affected` gate still skips it when nothing in `@opencode/app` or `@opencode/stories`, or their workspace dependencies, changed (`.github/workflows/check-test.yml`).
   - Each OS has the legs `app-1` and `app-2` (the shards), `session-ui-components` and `app-components`.
-  - The required `e2e (linux)` and `e2e (windows)` checks only aggregate those legs (`:329-352`).
+  - The `e2e (linux)` and `e2e (windows)` checks only aggregate those legs.
   - A pass on retry leaves a leg green. Check each Windows shard report's Flaky list, and make sure your new tests are not on it.
 - **Download:** `gh run download <run-id> -n playwright-windows-app-2-1 -D <dir>`. Artifacts are named `playwright-<os>-<leg>-<run attempt>` (`check-test.yml:314-327`). App e2e results are in `<dir>/app/e2e/`.
 - **In CI, the failing first attempt has no trace and no video** (`playwright.config.ts:62-65`).
diff --git a/.opencode/skills/component-tests/SKILL.md b/.opencode/skills/component-tests/SKILL.md
index 530a6166856..cea29788039 100644
--- a/.opencode/skills/component-tests/SKILL.md
+++ b/.opencode/skills/component-tests/SKILL.md
@@ -101,11 +101,11 @@ From the owning package (`packages/session-ui` or `packages/app`):
 5. **Windows.** Check platform keys and labels and font-dependent geometry. The Windows component legs are the authoritative check.
    - Do not skip on Windows, and do not raise `config.ts` timeouts, retries or the `mount` wait.
    - If one test does genuinely expensive work, call `story.setTimeout(...)` in that test, with a comment that names the work.
-6. **Confirm CI ran the component legs.** E2E never runs on pushes to dev (`check-test.yml:28`), so no post-merge run catches a regression.
-   - The `affected` gate filters on `@opencode/app` and `@opencode/stories` (`check-test.yml:60-61`). `ui`, `session-ui` and `gui-plugins` count through app's dependencies. `workflow_dispatch` forces the legs on (`check-test.yml:56-59`).
+6. **Confirm CI ran the component legs.** Pull requests run only the affected unit tests on Linux, so the component legs come from a push to dev or beta, a nightly run or the release gate.
+   - The `affected` gate filters on `@opencode/app` and `@opencode/stories`. `ui`, `session-ui` and `gui-plugins` count through app's dependencies. Manual runs force the legs on.
    - `check-test.yml` is listed in `globalDependencies` (`turbo.json:8`), so editing it marks every package affected.
-   - The required `e2e (linux)` and `e2e (windows)` checks also pass when no leg was needed (`check-test.yml:329-352`). Look for `e2e (<os>, session-ui-components)` and `e2e (<os>, app-components)`.
-   - If CI skipped them, report your local cold run in the PR.
+   - The `e2e (linux)` and `e2e (windows)` checks also pass when no leg was needed. Look for `e2e (<os>, session-ui-components)` and `e2e (<os>, app-components)`.
+   - If no push has run them, report your local cold run in the PR.
 
 ## When CI flakes