The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes data from 78 external APIs, creating a surface for indirect prompt injection.
Ingestion points: Data is fetched from 78 external sources, including NCBI, PubChem, and the World Bank.
Boundary markers: The instructions lack explicit delimiters to prevent the agent from interpreting data content as instructions.
Capability inventory: The skill uses tools such as WebFetch, curl, and Python execution to process data.
Sanitization: No sanitization is performed on data retrieved from the remote endpoints.
[COMMAND_EXECUTION]: The skill directs the agent to generate and execute local Python code for specific integration requirements.
Evidence: In references/brenda.md, the agent is instructed to write and run a Python script using the zeep library to interface with the BRENDA database's SOAP API, as REST is not supported.
[SAFE]: The skill demonstrates safe management of sensitive credentials.
Evidence: SKILL.md provides instructions for the agent to load API keys from the shell environment or a local .env file, which is a standard security best practice for secret management.

database-lookup