Security audit
Security checks for vulnerabilities and agentic risk
This writing helper reads a chosen document and creates a separate humanized output file, with no hidden network, credential, or persistence behavior found.
Before publishing or replacing original content, review the generated '-HUMAN.md' file and change log for meaning changes, factual accuracy, and unwanted tone shifts. Expect the skill to create local output files, but no hidden external access or privileged behavior was found.
The usage section says the skill processes a file and outputs a new -HUMAN.md file plus a change log, which is a user-data-affecting behavior. While the output is described, there is no explicit warning or caution about reviewing rewritten content before publishing or about the creation of additional files.
The skill explicitly states it will create a new "-HUMAN.md" file, which means it performs filesystem writes. While this appears to be part of the intended functionality rather than malicious behavior, the documentation does not clearly warn users up front that running the skill changes local state and creates output files, which can surprise users or cause unintended file proliferation or overwrites depending on implementation.
