aif-grounded
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill utilizes a context-loading mechanism that exposes the agent to potential malicious instructions embedded in the repository.
- Ingestion points: Step 0 of the workflow requires the agent to read additional instructions from
.ai-factory/skill-context/aif-grounded/SKILL.md. - Boundary markers: The skill lacks delimiters or instructions to ignore potentially harmful content within the context file. In fact, it explicitly instructs the agent that these external rules "win" and must be treated as project-level overrides.
- Capability inventory: The skill is granted access to sensitive tools including
Bash,Read,Write, andEdit, which could be triggered by instructions found in the context file. - Sanitization: No validation or sanitization process is defined for the content loaded from the context file.
Audit Metadata