The Agent Skills Directory

[PROMPT_INJECTION]: No evidence of prompt injection or instructions to bypass safety guidelines was found. The skill maintains a consistent expert persona focused on SEO without attempting to override system constraints.
[DATA_EXFILTRATION]: The skill does not perform network operations or access sensitive credentials. It references a local marketing context file (product-marketing.md), which is a standard practice for context-aware agents and does not pose an exfiltration risk.
[REMOTE_CODE_EXECUTION]: There are no commands or instructions that involve downloading or executing remote code or scripts.
[COMMAND_EXECUTION]: The skill is purely instructional and does not utilize shell commands or system-level execution tools.
[OBFUSCATION]: The content is written in clear, plain text. No hidden characters, Base64 encoding, or homoglyph-based obfuscation were detected.
[CREDENTIALS_UNSAFE]: No hardcoded secrets, API keys, or private keys are present in the provided files.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests external context from a product marketing file, it lacks the technical capabilities (such as shell access or network requests) that could be exploited via indirect injection. The risk remains negligible as the output is limited to text-based strategy recommendations.

programmatic-seo