The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes instructions to retrieve and analyze content from arbitrary external URLs using search and extraction tools. This untrusted data is then used to generate content outlines and analysis, creating a pathway for indirect prompt injection.
Ingestion points: The skill ingests data from external websites via tavily/extract and tavily/search-assistant tools as seen in the SERP Analysis Process section of SKILL.md.
Boundary markers: The provided prompt templates lack clear delimiters or instructions for the agent to distinguish between the skill's instructions and potentially malicious instructions contained within the extracted web content.
Capability inventory: The skill is authorized to use the belt CLI tool via Bash, which performs network operations to search and extract data.
Sanitization: No mechanisms are described for sanitizing or filtering instructions out of the data retrieved from external sources before it is processed by the model.
[EXTERNAL_DOWNLOADS]: The skill documentation suggests installing external CLI components and additional functional modules using npx and references configuration guides hosted on the inference-sh GitHub repository.

seo-content-brief