jinko-calibration-subsampling
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows established patterns for interacting with the Jinkŀ platform using its official Python SDK. All workflows described relate directly to the skill's stated purpose of virtual population subsampling.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data fetched from the Jinkŀ platform, which represents a standard vulnerability surface. 1. Ingestion points: Trial descriptors and result scalars fetched via the SDK in SKILL.md and multiple reference files. 2. Boundary markers: The instructions do not implement explicit delimiters or 'ignore embedded instructions' warnings for platform data. 3. Capability inventory: The skill can create and modify project items, generate Vpops, and execute SDK-provided CLI scripts (jinko.cli.*). 4. Sanitization: Input validation appears to rely on the SDK's internal type checking rather than explicit instruction-level sanitization.
Audit Metadata