The Agent Skills Directory

[NO_CODE]: The skill is composed entirely of informational Markdown documents (SKILL.md, reference.md, reference-documents.md, and reference-samples.md). It contains no Python scripts, Node.js packages, or shell commands.
[SAFE]: The contents of the skill were thoroughly audited for malicious patterns. No prompt injection, obfuscation, or data exfiltration attempts were found. The instructions are focused solely on literary analysis and screenwriting structural design.
[INDIRECT_PROMPT_INJECTION]: The skill guides the agent in analyzing external data such as user-provided script treatments and story outlines. While this processing of untrusted data theoretically presents an injection surface, the skill provides no technical capabilities (such as file writes or network access) for an attacker to exploit even if an injection occurred.

sw-series-engine-bible