Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed text-editing helper for making prose sound less AI-generated, with no executable code, persistence, credential access, or hidden behavior found.
Install only if you want a writing editor that may rewrite documents to remove AI-sounding style markers. Do not use it to bypass academic, workplace, platform, or compliance rules that require disclosure of AI assistance, and review file edits before accepting them.
The phrase "Or invoke directly when editing documents" is ambiguous about how and when the skill should activate. It does not define specific trigger phrases, constraints, or exclusion conditions, which could cause unintended invocation during normal document-editing workflows.
The skill’s stated purpose is to make text appear more human-written by removing signs of AI authorship, which can be used to conceal AI assistance or evade policies, reviewers, or provenance checks. The instructions go beyond ordinary copyediting by explicitly teaching stylistic laundering and adding 'personality' to mask machine-generated origin.
The manifest description says the skill will "Remove signs of AI-generated writing from text" and "Detects and fixes AI patterns," but it does not define when the skill should activate, what inputs qualify, or any exclusion conditions. In a manifest file, this kind of broad phrasing can overlap with many ordinary editing requests and may cause unintended invocation.
The skill is granted Write and Edit capabilities, but its metadata and user-facing framing do not clearly warn that it may directly modify files rather than just suggest edits. This can lead to unintended content changes, especially if a user invokes the skill in a repository or notes directory expecting analysis-only behavior.
No suspicious patterns detected.
