scientific-clarity-checker
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external scientific documents, which serves as a data ingestion surface for potentially untrusted content. A malicious document could contain embedded instructions meant to influence the agent's behavior. While the skill uses structured matrices and checklists to organize the review, it does not explicitly instruct the agent to ignore instructions found within the source material. However, the impact of such an injection is extremely limited as the skill has no defined tools or permissions for network activity, file system access, or command execution.
- Ingestion points: User-provided scientific documents as referenced in SKILL.md and resources/methodology.md.
- Boundary markers: Absent; the skill lacks specific delimiters or instructions to ignore commands within the ingested text.
- Capability inventory: No exploitable capabilities detected; no file-write, network, or subprocess tools are utilized by the skill scripts or metadata.
- Sanitization: Absent.
Audit Metadata