developmental-band-translator
Pass
Audited by Gen Agent Trust Hub on May 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates exclusively on curriculum metadata, facilitating the mapping of educational items (KUDs, Learning Targets) to developmental bands based on academic frameworks like Understanding by Design (UbD).
- [DATA_EXPOSURE]: No sensitive file paths, credentials, or network operations were detected. The skill processes input curriculum data and produces structured output with no external exfiltration points.
- [REMOTE_CODE_EXECUTION]: There are no patterns suggesting the download or execution of external scripts or packages. The logic is contained entirely within the prompt instructions.
- [PROMPT_INJECTION]: The skill includes strong negative constraints, such as the 'Source-voice preservation rule', which explicitly forbids the model from rewriting or paraphrasing source content, thereby mitigating risks of unintended instruction following or output manipulation.
- [INDIRECT_PROMPT_INJECTION]: While the skill ingests untrusted data (KUD and LT JSON files), it lacks any capabilities (such as shell access, file writes, or network requests) that could be exploited via an injection. The transformation is limited to structural tagging.
Audit Metadata