Security audit
Security checks for vulnerabilities and agentic risk
This skill is a straightforward text-polishing helper that can overwrite explicitly provided text files, so users should review changes but there is no evidence of hidden or malicious behavior.
Install only if you want an agent to rewrite prose in files you explicitly provide. Prefer asking for a diff or preview first, especially for source-adjacent files, documentation, or CLAUDE.md, because the skill is designed to write cleaned text back in place.
The skill advertises activation on broad phrases like 'clean up the copy' and related everyday editing requests, which can cause unintended invocation outside the user's intended workflow. Mis-triggering can lead to unsolicited rewriting of user content and, combined with write capability, increases the chance of accidental content modification.
The skill states it reads and rewrites files in place, and later says to write the cleaned version back, but it does not require user confirmation, backup creation, or preview before modification. This is dangerous because a mistaken invocation or over-aggressive rewrite can silently alter documentation or source-adjacent text, causing data loss or unwanted semantic changes.
