wp-wpcli-and-ops/examples/README.md

Disposable native WP-CLI diagnostic experiment

Run only on a disposable Linux Docker host that you control. The runner refuses existing fixture names, uses an internal network, publishes no ports, and creates WordPress/MariaDB data in tmpfs. It removes only containers/networks that it created. It performs synthetic setup writes; it never points at a real website or uses real credentials.

The two original helper sources, BB Skills adapted performance source, setup and runner are included. All experiment sources are GPL-2.0-or-later. The evidence records image IDs, digests, exact source hashes, WP-CLI checksum source and actual versions. No AI client, multisite migration, persistent cache, production performance, load test or measured speedup is covered.

To reproduce, review the runner and acquire the WordPress, MariaDB and Node images named in the evidence. Download WP-CLI 2.12.0 from the official release URL in native-cli-evidence.json, verify its SHA-512 against the pinned official Dockerfile in that record, and save it as wp-cli.phar beside the runner. Copy the wp_cli object from the evidence to a sibling wp-cli-source.json. The PHAR and Node executable are deliberately not redistributed in this skill package. Read all prerequisites before executing python3 run-cli-fixture.py.

The original reports demonstrate two limits in this environment: WP-CLI 2.12.0 --autoload=on omitted auto/auto-on values, and the upstream performance helper assumed default content/plugin directories. The BB Skills performance revision counts wp_load_alloptions() values and resolves WP_CONTENT_DIR / WP_PLUGIN_DIR through WordPress. File existence is still only a presence signal: our dummy files provide no working cache and are not active plugins.

Both upstream helpers can exit 0 when WP-CLI or the WordPress target is missing. Read the JSON readiness fields, nulls and failed subcommands. The helper has no built-in subprocess timeout; the runner applies an external limit. Review diagnostic paths and operational details before sharing output.