The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to extract product information such as names, prices, and images from external shop pages to be used in API calls or product feeds. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the external site's metadata or content to influence the agent's behavior.
Ingestion points: External product pages as described in SKILL.md.
Boundary markers: No specific delimiters or safety instructions are defined to isolate the extracted metadata from the agent's core instructions.
Capability inventory: The agent is authorized to perform API calls to the MeltFlex backend and write code directly into the user's project files.
Sanitization: While the skill demonstrates the use of URL encoding for address interpolation, it does not specify sanitization or validation rules for text-based metadata extracted from third-party sources.
[COMMAND_EXECUTION]: The skill contains several curl command examples for interacting with the MeltFlex API and performing diagnostic header checks. These are standard developer operations for the described integration and do not involve unsafe shell piping or execution of untrusted remote scripts.
[DYNAMIC_EXECUTION]: The skill involves the generation of framework-specific code templates (React, Vue) and standard HTML iframes to be integrated into a user's codebase. This is a functional requirement for providing integration guidance and uses the static templates provided in the documentation.

meltflex-embed