inquiring
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources to validate claims, creating an injection surface.
- Ingestion points: External content is fetched from public and primary sources via the
/researchtool as specified inSKILL.md. - Boundary markers: Absent. There are no instructions to use delimiters or ignore instructions embedded within the fetched evidence.
- Capability inventory: According to
SKILL.md, the agent can write markdown logs, generate HTML views, and create/execute Python code within Jupyter notebooks (findings/<slug>.ipynb). - Sanitization: Absent. The instructions do not require sanitizing external data before it is recorded in the findings log or used in code generation.
- [DYNAMIC_EXECUTION]: The skill directs the agent to dynamically generate and execute Python code in a Jupyter notebook (
findings/<slug>.ipynb) when a claim requires computational evidence. This code is constructed using information gathered from external, potentially untrusted sources.
Audit Metadata