The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documents procedures for querying external WMS OData entities, which establishes a surface for indirect prompt injection from the retrieved data fields.
Ingestion points: External data is fetched from entities like ArchivedShippingLicensePlateContents, Tasks, and Materials as outlined in references/archived-shipping-lpc.md and references/receiving-tasks.md.
Boundary markers: The instructions do not recommend the use of delimiters or specific warnings to ignore instructions that might be embedded within data returned from the WMS (e.g., in Material descriptions or Task notes).
Capability inventory: The skill encourages the use of dxs odata execute and $api calls to retrieve content which is then used for reporting and logical calculations (e.g., weight functions in references/weight-and-units.md).
Sanitization: There is no guidance provided for sanitizing or escaping the data retrieved from the external service before it is processed or presented to the user.

footprint-entity-expert