The Agent Skills Directory

[SAFE]: The skill consists of Markdown instructions and interface configuration for planning video trailers. No malicious command execution, network exfiltration, or credential harvesting patterns were found during the analysis. The skill relies on standard agent interactions without invoking external scripts.
[NO_CODE]: The skill files contain instructions and templates but do not include any functional code in Python, JavaScript, Shell, or other programming languages. It specifies a workflow rather than providing automated tools.
[INDIRECT_PROMPT_INJECTION]: While the skill involves researching external sources like song lyrics and trailer videos, it lacks exploitable capabilities (such as file-writing or subprocess execution) that could be triggered by malicious data. Furthermore, the instructions include proactive measures such as paraphrasing external content, limiting direct quotes, and verifying timestamps, which significantly reduces the risk of accidental execution of instructions embedded in external research data.

trailer-direction