Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@cypress.config.ts`:
- Around line 39-42: gatherClusterArtifacts currently writes all artifacts into
a single clusterDir under ARTIFACTS_DIR which causes later specs to overwrite
earlier data; update gatherClusterArtifacts to accept the current spec
identifier (e.g., specName or specPath) and create a per-spec subdirectory
(e.g., path.join(ARTIFACTS_DIR, 'cluster', sanitize(specName))) before creating
podLogsDir so pods.yaml/services.yaml and logs are written per-spec, and update
all callers of gatherClusterArtifacts (the after-each hook) to pass the spec
name/path; ensure the spec string is sanitized to form a safe filesystem name.
- Line 3: Replace the shell-based execSync usage with execFileSync to avoid
shell interpolation: import execFileSync from 'child_process' (replace
execSync), change the runOC helper to accept an argv array (e.g. args: string[])
instead of a single interpolated string, and call execFileSync('oc', argsArray,
{ env: {..., KUBECONFIG: kubeconfigPath}, encoding: 'utf-8' }) so
KUBECONFIG_PATH and values like OLS_NAMESPACE are passed as discrete argv
entries rather than concatenated into a shell command; update all call sites
(lines ~26-35) to use array-based arguments like ['get','<resource>','-n',
OLS_NAMESPACE, '-o','yaml'].