wordpress-admin

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill embeds plaintext credentials (local admin/admin123) and an lftp command containing a raw FTP password, and it shows places where API keys/authorization headers must be inserted, which forces the LLM to handle or could cause it to reproduce secret values verbatim.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow (e.g., calling scripts/wp-page.py/scripts/wp-seo.py) executes and forwards user-provided CLI text like --content, --title, --focus-kw, and --meta-desc into WP-CLI command strings (via subprocess.run(..., shell=True)), so outsider-authored free text can be ingested as arguments without selecting a specific trusted item first.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I scanned the documentation and scripts for literal, high-entropy credentials. The local Docker admin credential "admin / admin123" (SKILL.md line 30) is a low-entropy setup/example password and is ignored per policy. API key placeholders like "API_KEY" and instructions to store keys in hidden files (e.g., /root/.pexels-api-key) are documentation placeholders or guidance and are ignored.

However, the lftp upload command in SKILL.md contains a quoted, random-looking password '@#s;v1#%1M$+' embedded directly in the command line. That is a high-entropy literal value that would grant FTP access if valid, so I flag it as a real secret.