The Agent Skills Directory

[SAFE]: The skill operates entirely within the context of text analysis and classification for medical research. It does not utilize any external tools, network operations, or file system modifications.
[SAFE]: Hard rules (9 and 10) explicitly prohibit the fabrication of metadata, references, or study features, which mitigates risks related to hallucination or deceptive data generation.
[SAFE]: While the skill ingests untrusted data (medical papers/abstracts) as its primary input, which represents a surface for Indirect Prompt Injection (Category 8), the agent lacks any dangerous tools or permissions (e.g., shell access, network requests, file writes) that could be exploited by such an injection. The impact is limited to the accuracy of the classification output itself.
[SAFE]: All external references and metadata (such as the GitHub repository link) are consistent with the provided author context (AIPOCH) and represent legitimate vendor resources.

study-design-identifier