The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data sources and source material, creating a potential attack surface for indirect prompt injection where malicious instructions could be embedded in the data processed by the agent.
Ingestion points: The skill instructions in SKILL.md and references/sandbase-api-map.md explicitly mention reading "source material" and "external data source" to fulfill diagram generation tasks.
Boundary markers: The provided instructions do not specify the use of delimiters or "ignore embedded instructions" markers when handling these external data inputs.
Capability inventory: The skill utilizes several tools including sandbase_run for executing external capabilities and sandbase_inspect for reading schemas, providing a functional path for injected instructions to interact with external systems.
Sanitization: There are no explicit requirements or steps provided for sanitizing or escaping the content retrieved from external sources before it is processed by the agent.

excalidraw-diagram-generator