place-based-curriculum-orchestrator
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The orchestrator takes several user-provided inputs (e.g., local place descriptions, community knowledge) and directly places them into the prompt for the model. This configuration creates a potential surface for indirect prompt injection, although the risk is minimized by the skill's lack of executable tools or network access. \n- Ingestion points: Variable inputs starting_point, learner_stage, subject_or_curriculum_content, local_place, community_knowledge_holders, and constraints in SKILL.md. \n- Boundary markers: User inputs are inserted using simple curly-brace placeholders without specific delimiters or isolation instructions for the model. \n- Capability inventory: The skill is restricted to text generation and orchestration of other educational skills; it contains no tools for system modification, file access, or network communication. \n- Sanitization: No sanitization or input validation logic is present to filter malicious or malformed user data.
Audit Metadata