The Agent Skills Directory

[NO_CODE]: The skill is entirely documentation-based, consisting of Markdown files with prompt templates and research. There are no executable scripts or binaries.
[SAFE]: No network, file system, or credential access patterns were found. The skill operates purely as a text advisor.
[PROMPT_INJECTION]: The skill includes instructions for analyzing and optimizing user-provided prompts, including guidelines for detecting adversarial inputs and testing for jailbreaks. It uses structural delimiters like XML tags which are a security best practice. While it processes untrusted user data (prompts), it has no access to dangerous tools or capabilities, neutralizing the risk of indirect prompt injection. (Ingestion: user prompts in chat; Boundaries: XML tags recommended in templates; Capabilities: none detected; Sanitization: none).

prompt-engineering