aurelius-mapping
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of technical documentation and instructional guidelines for using the TMS Aurelius ORM framework in Delphi. The recommendations, such as using
RegisterEntityin the initialization section to prevent RTTI stripping by the linker, are standard best practices for this framework. - [OBFUSCATION]: The documentation uses zero-width space characters (e.g.,
​) within long technical identifiers likeAurelius.Mapping.Attributes. This is a common formatting technique in documentation to allow for proper text wrapping and does not hide malicious commands or redirect the agent's behavior. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied class definitions or mapping requirements to generate Delphi code. This is a standard functional surface for a mapping skill. The provided instructions constrain the agent to follow specific ORM patterns, which mitigates the risk of the agent generating unintended or unsafe code beyond the requested mapping logic.
- Ingestion points: User requests for creating or fixing Delphi entity mappings in
SKILL.mdandreferences/mapping.md. - Boundary markers: The skill uses clear headings and code blocks to separate instructions from reference material.
- Capability inventory: The skill is limited to code generation and explanation; it does not involve subprocess execution, file writes, or network requests.
- Sanitization: Not applicable as the output is static source code intended for manual integration by the developer.
Audit Metadata