The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references downloading standard academic datasets and lexicons, such as the NRC Emotion Lexicon from saifmohammad.com and GloVe embeddings from Stanford University's NLP project. It also utilizes standard pre-trained models from the Hugging Face ecosystem. These are well-known and trusted sources in the data science community.
[COMMAND_EXECUTION]: The skill provides instructions for setting up environments using standard package managers (pip for Python, install.packages for R) and recommends using Git for version control to track analytical progress. These operations are aligned with the skill's primary purpose of providing a reproducible research environment.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external text corpora provided by the user. While this creates a potential surface for indirect prompt injection, the skill's instructions focus on statistical analysis and human validation of outputs rather than automated decision-making or sensitive actions based on document content. The documentation strongly emphasizes 'human validation and multiple diagnostics' as essential steps, which serves as a primary mitigation against the effects of potential injection in the analyzed data.

text-analyst