The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external search providers (Tavily and Exa) to research industry contexts and competitor landscapes. If these external sources contain malicious instructions, they could influence the agent's behavior during case study generation.
Ingestion points: Outputs from tavily/search-assistant, exa/search, and exa/answer are processed by the LLM (SKILL.md).
Boundary markers: The skill does not provide specific delimiters or instructions to ignore potential commands within the ingested search data.
Capability inventory: The skill can execute Python code (infsh/python-executor), write files (plt.savefig), and perform further network searches.
Sanitization: No evidence of sanitization or validation for the content returned by external search tools.
[DYNAMIC_EXECUTION]: The skill utilizes a tool (infsh/python-executor) to run Python code at runtime for generating data visualization charts. While the included example is restricted to using matplotlib for generating a .png file, the capability allows for dynamic assembly and execution of code.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the belt CLI and additional skills from the inference-sh and belt-sh organizations on GitHub and NPM. These resources are standard components of the underlying platform ecosystem.
[COMMAND_EXECUTION]: The skill executes various CLI commands via the belt tool, including login operations and running specific applications like search assistants and code executors.

case-study-writing