duck-teach
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted data from the codebase and user input to generate tutorials.
- Ingestion points: Data enters the agent context through user queries (e.g., "explain this") and automated codebase searches for project-specific patterns as defined in
SKILL.md. - Boundary markers: The skill provides explicit instructions to preserve core safeguards and references a
references/GUARDRAILS.mdfile that requires safety carve-outs and bounded approval for any actions. - Capability inventory: The skill has the capability to read workspace files via codebase search and display content to the user.
- Sanitization: Instructions in
SKILL.mdmandate refusing any recommendation that would weaken trust boundaries, data-loss prevention, or security controls, requiring a safe alternative instead. - [SAFE]: The skill explicitly defines boundaries that prevent the execution of edits or commands without user approval and handles complex scenarios by handing off to specific debug or review modules, maintaining a controlled execution environment.
Audit Metadata