The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified in the skill's instructions or scripts. The code behavior is consistent with its stated purpose of scientific data analysis.- [EXTERNAL_DOWNLOADS]: The skill relies on standard, well-known Python packages listed in requirements.txt (including numpy, pandas, scipy, and scikit-bio). These are reputable libraries from established ecosystems. No remote script execution or binary downloads from untrusted sources are performed at runtime.- [COMMAND_EXECUTION]: The primary script (scripts/main.py) performs data processing and file I/O using standard Python libraries. It does not utilize dangerous functions such as os.system(), subprocess.run(), or eval() to execute shell commands or dynamic code.- [DATA_EXFILTRATION]: No network-enabled operations (e.g., requests, curl, urllib) were found in the provided code. The script's operations are limited to reading user-provided TSV files and writing the resulting analysis artifacts to local paths specified by the user.- [PROMPT_INJECTION]: The SKILL.md instructions follow standard instructional patterns for an AI agent without attempting to override safety filters, bypass guidelines, or extract system prompts.- [INDIRECT_PROMPT_INJECTION]: The script processes external data from TSV files and incorporates sample names and metadata into HTML/Markdown reports. While the script does not sanitize this data before interpolation into the HTML template, this represents a low-risk surface for typical data reporting tools and does not provide an execution path for the agent.

microbiome-diversity-reporter