pp-wordpress

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill's recall/playbook JSON can include a "token" in slots_resolved and the agent is instructed to substitute slots_resolved entries into playbook steps (plus it tells users to set WORDPRESS_USER / WORDPRESS_APP_PASSWORD), which requires reading and inserting secret credentials verbatim into commands or outputs.

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

• Third-party content exposure detected (low risk: 0.10). SKILL runtime path ingests only user-provided natural-language questions to select/parameterize the wordpress-pp-cli commands (e.g., via recall/command resolution) and then reads first-party WordPress REST content from a user-specified site using Application Passwords, not from any outsider-controlled feed/queue that the workflow monitors by default.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill document includes installation commands that may be executed at runtime (e.g., an "install" flow) which fetch and execute remote code from GitHub module paths such as github.com/mvanhorn/printing-press-library/library/developer-tools/wordpress/cmd/wordpress-pp-cli@latest.