skills/ndpvt-web/arxiv-claude-skills/deltaevolve-accelerating-scientific-discovery/Gen Agent Trust Hub
deltaevolve-accelerating-scientific-discovery
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill involves a workflow where the agent generates executable code snippets and then runs them through a deterministic evaluation function to calculate a score. This iterative execution is the primary intended purpose of the skill for discovering optimized algorithms.
- [PROMPT_INJECTION]: The skill utilizes a 'progressive disclosure' mechanism where previous code generations and their associated performance deltas are interpolated back into the prompt for subsequent iterations. This creates an indirect prompt injection surface, though it is mitigated by the use of structured delimiters (e.g.,
#Code-Start,#Delta-Summary-Start) and is a core component of the iterative learning process. - [EXTERNAL_DOWNLOADS]: The skill references an external research paper on Arxiv (arxiv.org), which is a well-known and trusted academic repository.
Audit Metadata