thinking-science
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external sources to support the experimental phase of the scientific method. This data is then used in a context where potentially dangerous tools are active.
- Ingestion points: The skill utilizes WebSearch and WebFetch to gather evidence and data as described in the Experiment Design phase in SKILL.md.
- Boundary markers: The instructions lack specific delimiters or prompt-based instructions to ensure the agent ignores malicious directives potentially embedded in the fetched external content.
- Capability inventory: The skill is configured with access to the Bash tool (command line execution) and Write tool (file system modification), which could be targets for instructions found in ingested data.
- Sanitization: No explicit sanitization, filtering, or validation steps are defined for the content retrieved from external URLs before it is processed by the agent.
Audit Metadata