The Agent Skills Directory

[SAFE]: No malicious patterns, prompt injections, or obfuscated content were detected. The skill is entirely instructional and adheres to the stated purpose of helping users optimize their content for AI systems.
[NO_CODE]: The skill consists exclusively of markdown documentation and a JSON evaluation file. There are no Python scripts, JavaScript files, or shell commands present, which eliminates the risk of code-based attacks.
[EXTERNAL_DOWNLOADS]: The skill references several external resources, including official Google Search documentation, GitHub repositories under the GoogleCloudPlatform organization, and well-known industry tools such as Semrush, Ahrefs, and Perplexity. All referenced links are for informational purposes and target established, trusted domains.
[PROMPT_INJECTION]: Analysis of the skill's instructions shows a standard indirect prompt injection surface where the agent reads local project context (e.g., .agents/product-marketing.md) and user-provided website data. However, the skill lacks dangerous capabilities such as subprocess execution, file-system writes, or network exfiltration tools, making this ingestion surface safe in the current implementation.
Ingestion points: Reads project context from .agents/product-marketing.md (SKILL.md) and processes user-supplied brand/query data.
Boundary markers: Absent.
Capability inventory: The skill utilizes analytical SEO tools for data retrieval and content assessment; no direct system command execution or dynamic code evaluation is present.
Sanitization: Not explicitly implemented.

ai-seo