The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the denario[app] package from a public registry using pip or uv during the setup process.
[COMMAND_EXECUTION]: The skill provides a command-line interface denario run which executes a local web server to host a graphical user interface for research management.
[REMOTE_CODE_EXECUTION]: The core functionality includes a get_results() method that executes computational experiments and generates visualizations based on AI-developed methodologies. This represents a dynamic code execution environment where agents generate and run scripts locally to produce research findings.
[PROMPT_INJECTION]: Indirect Prompt Injection vulnerability surface detected via the ingestion of untrusted research data.
Ingestion points: Data enter the agent context through set_data_description, set_idea, and set_method functions in SKILL.md.
Boundary markers: There are no documented delimiters or instructions to the agent to ignore potentially malicious embedded instructions within the research datasets or methodologies.
Capability inventory: The skill possesses capabilities to perform file system writes (get_paper generating LaTeX source) and execute generated code for computational analysis (get_results).
Sanitization: No evidence of input validation, escaping, or filtering of external content is provided in the skill definitions.

denario