The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected within the skill instructions.
[NO_CODE]: The skill contains only documentation and procedural guidelines in markdown format. It does not include scripts, binaries, or automation code.
[PROMPT_INJECTION]: The instructions do not contain attempts to bypass safety guardrails, override system prompts, or manipulate agent behavior beyond the intended audit scope.
[INDIRECT_PROMPT_INJECTION]: Although the skill involves analyzing external data (screenshots/implementations), it lacks the capabilities (such as shell execution, file system writes, or network access) required to act on instructions that might be embedded in that data.
[DATA_EXFILTRATION]: There are no commands or instructions that could result in the unauthorized access or transmission of sensitive data.

visual-audit