assessment-validity-checker
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill accepts multiple untrusted inputs (such as assessment_description and intended_learning) and interpolates them directly into the core prompt. It does not use delimiters (e.g., XML tags, triple quotes) or specific instructions to ignore potential commands embedded within the user-provided assessment data. This creates a surface where an attacker could provide an assessment description containing malicious instructions to override the agent's behavior.
- Ingestion points: The prompt in
SKILL.mdinterpolates {{assessment_description}}, {{intended_learning}}, {{student_level}}, {{subject_area}}, {{assessment_purpose}}, {{marking_approach}}, and {{stakes}}. - Boundary markers: Absent; user input is mixed directly with system instructions.
- Capability inventory: The skill primarily performs text analysis and does not explicitly invoke external tools, network operations, or file system modifications, which limits the potential impact of an injection.
- Sanitization: No input validation or escaping is applied to the data before interpolation.
Audit Metadata