The Agent Skills Directory

[SAFE]: No malicious patterns were detected. The skill is instructional and focuses on data architecture review logic. It does not contain obfuscated code, unauthorized network calls, or persistence mechanisms.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data warehouse metadata, which creates a surface for indirect prompt injection. 1. Ingestion points: Table schemas, database lists, and metadata provided by the user in the initial step of the audit process. 2. Boundary markers: Absent; there are no instructions to use delimiters or ignore embedded instructions within the processed metadata. 3. Capability inventory: The skill is restricted to generating reports and recommendations; it lacks tools for subprocess execution, file writes, or network requests. 4. Sanitization: Absent; no filtering or escaping of user-provided metadata is defined. The risk is considered safe due to the absence of exploitable capabilities.

data-solid