probability-theory-expert
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary purpose is instructional and mathematical. It does not contain any malicious code, hidden instructions, or attempts to bypass safety filters.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an ingestion surface for untrusted external data as part of its modular update workflow.
- Ingestion points: Instructions in
modules/research-checklist.mddirect the agent to fetch information from external web sources (Wikipedia, ArXiv) using platform tools (adn_skills). - Boundary markers: The instructions lack explicit boundary markers or warnings to ignore potentially malicious instructions embedded in the retrieved research data.
- Capability inventory: The agent is instructed to read external content and then write or modify its own internal modules (e.g.,
modules/core-guidance.md) and metadata. - Sanitization: There are no instructions for validating, filtering, or sanitizing the retrieved external content before it is integrated into the skill's knowledge base.
Audit Metadata