cognitive-scaffolding
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists primarily of documentation and illustrative code snippets in TypeScript and Python intended to help developers structure prompts. It does not include functional scripts or executables.
- [INDIRECT_PROMPT_INJECTION]: The code examples describe how to build prompts using external data and user tasks. The skill explicitly recommends using boundary markers, such as XML-style tags (
<section>,<reference>,<current_task>), which is a best practice for preventing the agent from misinterpreting data as instructions.
Audit Metadata