The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions are focused on content creation and do not include any commands to bypass safety filters or ignore previous instructions. The language used is purely instructional for the intended task.\n- [DATA_EXPOSURE]: The skill accesses local marketing context files (e.g., .agents/product-marketing.md) to personalize generated content. This is standard behavior for providing business context and does not involve accessing or exfiltrating sensitive system data or credentials.\n- [REMOTE_CODE_EXECUTION]: The skill does not perform any network requests to download or execute external code, nor does it define any package dependencies.\n- [OBFUSCATION]: All instructions and reference files are in plain text with no hidden or encoded elements, homoglyphs, or zero-width characters.\n- [DYNAMIC_CONTEXT_INJECTION]: The skill does not use shell execution placeholders or dynamic command injection patterns in its documentation.

competitors