The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes BCF files, which are external data sources that could theoretically contain malicious instructions designed to influence the agent's behavior during data processing.\n
Ingestion points: The bcfTopics.load(data) method in SKILL.md and references/methods.md is the primary entry point for external binary data.\n
Boundary markers: The component includes a strict mode in its configuration to validate imported data properties against allowed sets, providing a layer of control over incoming metadata.\n
Capability inventory: The skill's capabilities are restricted to BIM data management (creating topics, capturing 3D views) and local file export via Blob objects. It does not possess capabilities for arbitrary command execution or unauthorized network communication.\n
Sanitization: The implementation relies on established industry-standard libraries, specifically jszip for container handling and fast-xml-parser for XML processing, to ensure structured data is handled correctly.\n- [SAFE]: The skill demonstrates secure development practices, including explicit resource disposal patterns and the use of GUIDs for cross-referencing objects to prevent memory leaks. All external dependencies are well-known, legitimate packages used within the intended architectural domain.

thatopen-impl-bcf