palantir-data-handling
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were detected. The skill provides boilerplate code for Palantir Foundry that adheres to data privacy best practices, including hashing sensitive identifiers and implementing access control markings. The provided PySpark transforms demonstrate security best practices such as hashing PII (email), masking data (phone numbers), and implementing data retention policies.\n- [INDIRECT_PROMPT_INJECTION]: While the skill defines workflows that ingest data from external datasets (e.g., /Company/datasets/raw_customers), this is the primary purpose of the skill. The instructions explicitly include sanitization steps (hashing, redaction) to mitigate data exposure risks.\n
- Ingestion points: Input datasets such as /Company/datasets/raw_customers identified in SKILL.md.\n
- Boundary markers: The skill implements functional boundaries via redaction and hashing transforms.\n
- Capability inventory: File reading/writing via Foundry transforms and execution of Ontology Actions.\n
- Sanitization: Demonstrates robust sanitization using SHA-256 hashing and literal masking.
Audit Metadata