The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of data from external sources during app migration, which creates a potential surface for indirect prompt injection if the imported content contains malicious instructions designed to subvert agent logic.\n
Ingestion points: Data is received via the importResources function in references/appmigrationkit-patterns.md.\n
Boundary markers: The code includes validateManifest and version checks (sourceVersion) to verify incoming data.\n
Capability inventory: The skill uses FileManager to perform file system operations (write/delete) within the application's sandbox and app groups.\n
Sanitization: Implementation patterns suggest manifest validation and structural remapping of imported files.\n- [EXTERNAL_DOWNLOADS]: The skill contains references to documentation and API specifications hosted on an external domain that is not recognized as a trusted or well-known service provider.\n
Evidence: SKILL.md contains several links to https://sosumi.ai/documentation/.

appmigrationkit