The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the hypogenic and s2orc-doc2json packages from PyPI and clone external datasets from https://github.com/ChicagoHAI/HypoGeniC-datasets.git and https://github.com/ChicagoHAI/Hypothesis-agent-datasets.git.
[COMMAND_EXECUTION]: Documentation for the skill includes several shell commands for environment setup, including uv pip install, git clone, and the execution of local shell scripts such as bash ./modules/setup_grobid.sh and bash ./modules/run_grobid.sh.
[PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by ingesting external data sources (research PDFs and JSON datasets) which are interpolated into LLM prompts for hypothesis generation and inference.
Ingestion points: Data enters the context through files specified in the config.yaml (e.g., train_data_path) and literature PDFs stored in the literature/ directory.
Boundary markers: The prompt templates in references/config_template.yaml use curly-brace placeholders like {sample_text} and {data_samples} without explicit delimiters or safety instructions to distinguish data from instructions.
Capability inventory: The skill performs LLM inference and local execution of research tools and setup scripts.
Sanitization: No sanitization or content filtering of the external files was found in the provided configuration.

hypogenic