The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the installation of the openfda-pp-cli binary using npx from the @mvanhorn npm scope and go install from the vendor's GitHub repository (github.com/mvanhorn/printing-press-library). These are vendor-controlled resources used for setting up the required tooling.
[DATA_EXFILTRATION]: The skill documents a --deliver webhook:<url> flag that allows command output to be POSTed to an arbitrary external URL. This provides a direct mechanism for the agent to send processed data to external servers, which could be leveraged for data exfiltration if the agent is directed to a malicious endpoint.
[COMMAND_EXECUTION]: The primary function of the skill is to invoke the openfda-pp-cli binary with user-supplied arguments to query various FDA databases. It also includes an openfda-pp-cli which command that processes natural language queries to resolve command names.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from the OpenFDA API, which is an external data source.
Ingestion points: Data is ingested through various subcommands (e.g., animal-events, drug-events, food-recalls) that fetch records from the FDA.
Boundary markers: Responses are wrapped in a JSON envelope containing meta and results fields, which provides structure but does not include explicit instructions for the agent to ignore potential commands within the data.
Capability inventory: The skill possesses the ability to write to the local filesystem and perform network POST requests via the --deliver sink options.
Sanitization: No specific sanitization or escaping of the external API content is mentioned before it is processed by the agent.

pp-openfda