docs(fetcher): drop register words

docs/cloud-setup/bitwarden.md

@@ -148,7 +148,7 @@ The health check is the OAuth2 token exchange: a minted token proves the
 client, and no data endpoint is touched. A refused exchange is a health error
 carrying the response.
 
-End-to-end smoke tests live in `crates/fetcher/tests/e2e/smoke_remote.rs`
+Smoke tests against a live Bitwarden organisation live in `crates/fetcher/tests/e2e/smoke_remote.rs`
 and are `#[ignore]`d by default. They read `BITWARDEN_CLIENT_ID` and
 `BITWARDEN_CLIENT_SECRET` (required), plus `BITWARDEN_API_BASE` and
 `BITWARDEN_IDENTITY_URL` (optional region/self-host overrides), from

docs/cloud-setup/onepassword.md

@@ -150,7 +150,7 @@ the bearer JWT string. The only configurable service-side key is `limit`
 The profile's probe calls `/api/auth/introspect` with the token (this does
 not consume any per-endpoint event budget) and is healthy on a 2xx.
 
-End-to-end smoke tests live in `crates/fetcher/tests/e2e/smoke_remote.rs`
+Smoke tests against a live 1Password account live in `crates/fetcher/tests/e2e/smoke_remote.rs`
 and are `#[ignore]`d by default. They read `ONEPASSWORD_EVENTS_TOKEN`
 (required) and `ONEPASSWORD_API_BASE` (optional regional override) from
 `.env-cloud` (or `.env`):

docs/cloud-setup/slack.md

@@ -75,8 +75,8 @@ API calls.
    Go to **Manage Distribution**. Confirm all four checklist sections show
    green checkmarks, tick **I've reviewed and removed any hard-coded
    information**, and select **Activate Public Distribution**. This does not
-   list the app publicly - it is the technical prerequisite that unlocks
-   org-wide (rather than single-workspace) installation.
+   list the app publicly. Slack requires it before the app can be installed
+   org-wide rather than in a single workspace.
 
 5. **Install at the organization level**
 
@@ -145,7 +145,7 @@ The profile's probe calls `/api/auth.test` with the token; Slack answers 200
 either way, so `ok: false` is the refusal and is reported as a health error
 carrying Slack's reason.
 
-End-to-end smoke tests live in `crates/fetcher/tests/e2e/smoke_remote.rs`
+Smoke tests against a live Slack Enterprise org live in `crates/fetcher/tests/e2e/smoke_remote.rs`
 and are `#[ignore]`d by default. They read `SLACK_AUDIT_TOKEN` (required)
 from `.env-cloud` (or `.env`):
 

docs/testing.md

@@ -17,7 +17,7 @@ The general rules are hyperi-ai's testing standard (`standards/universal/testing
 
 Core holds the unit tests, the fixture and wiremock suites, and every integration test against a real service: Kafka, the Kafka TLS round trip, OpenBao, LocalStack, ClickHouse and MongoDB.
 
-Full adds the container extractor tests, the Kafka end-to-end tests and the batcher measurement, which moves 100k rows through real Kafka. `test.full.rust.filter` in `.hyperi-ci.yaml` leaves out `smoke_remote` and `runzero`, which need tenant credentials no runner holds. Run those by hand with the credentials their `#[ignore]` reasons name.
+Full adds the container extractor tests, the Kafka produce-and-consume tests in the `e2e` target and the batcher measurement, which moves 100k rows through real Kafka. `test.full.rust.filter` in `.hyperi-ci.yaml` leaves out `smoke_remote` and `runzero`, which need tenant credentials no runner holds. Run those by hand with the credentials their `#[ignore]` reasons name.
 
 The `db-odbc` suites build only with that feature and need the `odbc-postgresql` and `odbc-mariadb` drivers, so CI does not run them. Each starts its own database container through testcontainers.