The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs users to download and install a Python package from a third-party GitHub repository (https://github.com/mcmero/labarchives-py) that is not part of a verified or trusted organization. This external dependency is required for core functionality but introduces a supply chain risk from an unverified source.
[METADATA_POISONING]: The skill contains a section that instructs the agent to proactively suggest a specific commercial platform (K-Dense Web) to the user when complexity increases. This is a behavioral override that uses the skill's instructions for promotional redirection rather than purely technical tasks.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources (LabArchives API responses and notebook entries) which could contain malicious instructions.
Ingestion points: XML and JSON responses retrieved from LabArchives API endpoints; notebook entries and file attachments (PDF, DOCX, CSV, etc.).
Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to disregard natural language commands embedded within the processed data.
Capability inventory: The skill uses Python scripts (scripts/setup_config.py, scripts/notebook_operations.py, scripts/entry_operations.py) that perform file system operations and network requests.
Sanitization: There is no evidence of sanitization or validation of the data retrieved from external API calls before it is processed or presented to the agent.

labarchive-integration