The Agent Skills Directory

[PROMPT_INJECTION]: The skill does not contain any instructions that attempt to bypass safety filters, override agent behavior, or extract system prompts. The instructions are strictly focused on biological research methodology.
[DATA_EXFILTRATION]: No hardcoded credentials, sensitive file path access, or unauthorized network operations were identified. The network calls described are limited to well-known bioinformatics APIs (EBI MGnify, EuropePMC, etc.) required for the skill's stated purpose.
[OBFUSCATION]: The skill's content is entirely in plain text with no evidence of Base64 encoding, zero-width characters, homoglyphs, or other obfuscation techniques.
[EXTERNAL_DOWNLOADS]: The skill references a 'tooluniverse' Python library. This library is central to the skill's framework and is treated as a legitimate dependency for the research tools provided.
[REMOTE_CODE_EXECUTION]: There are no patterns suggesting the download and execution of remote scripts (e.g., curl to bash) or the use of dynamic execution functions like eval() or exec() with untrusted data.
[COMMAND_EXECUTION]: The skill does not perform arbitrary shell command execution or utilize system-level tools that could lead to privilege escalation.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as scientific abstracts and taxonomic profiles. While this presents an ingestion surface for untrusted data, the skill lacks high-privilege capabilities (like file writes or shell access) that would make such an injection exploitable. The risk is considered minimal and inherent to the data-retrieval task.

tooluniverse-microbiome-research