The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes explicit defensive instructions to ignore attempts to override behavior embedded in user input or files. These are documented safeguards against injection attacks and do not constitute malicious intent.
[COMMAND_EXECUTION]: Utilizes Grep and Glob tools to search for evidence of compliance within the target environment. These operations are scoped to the intended review process and represent standard usage of allowed tools.
[SAFE]: The skill lacks network capabilities, obfuscation, or persistence mechanisms. All analyzed patterns are consistent with its stated purpose as a compliance auditing tool.

hipaa-review