Security audit
typecho-blog-publish
Security checks for vulnerabilities and agentic risk
Overview
The skill's code and instructions largely match a Typecho publishing tool, but its runtime file-reading behavior (searching .env in parent/workspace/home paths) and included management/cleanup scripts broaden its access to workspace-wide secrets and blog content — review before use.
This skill appears to implement what it claims (publishing to Typecho) but has behaviors you should verify before installing: 1) Inspect load_env implementations (in scripts) — they search parent directories and ~/.openclaw/workspace/.env. If you have other secrets in ancestor .env files, the skill can read them. 2) Put the BLOG_PASSWORD into a dedicated .env inside an isolated skill workspace, and set file permissions to 600. Prefer an application-specific (limited) password on the blog account rather than a full admin password if possible. 3) Review scripts cleanup_blog.py and manage.py: they can edit/soft-delete posts; test them on a staging blog, not production. 4) Check publish_log.txt and any logging to ensure no passwords or sensitive fields are logged; remove/rotate secrets if they were logged during testing. 5) Audit network endpoints the skill contacts (default blog URL, any Unsplash/image download code) to confirm you trust them. 6) Run the skill in an isolated environment (or container) first, and consider limiting filesystem scope to the skill directory. If you need, ask the author to limit .env lookup to the skill directory only.
Static analysis
No suspicious patterns detected.