The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the infsh CLI tool to run specific applications for search and extraction, such as tavily/search-assistant and exa/extract.- [EXTERNAL_DOWNLOADS]: Links to external documentation and resources on the inference.sh domain are provided, along with installation commands using npx.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it fetches untrusted data from the internet and passes it directly into LLM prompts for analysis.
Ingestion points: Untrusted web content and search results retrieved via tavily and exa tools (SKILL.md).
Boundary markers: Basic tags like <search-results> are used in workflow examples, but there is no instruction to the AI to ignore instructions contained within that data.
Capability inventory: The skill can execute CLI commands, write search results to local files, and initiate network calls to LLM providers.
Sanitization: No sanitization or content filtering is implemented for the data extracted from the web before it is included in a prompt.

web-search