grug-brain-development
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains a section titled 'CRITICAL: Grug-Speak Mode' in
SKILL.mdwhich uses strong override language to force the agent into a specific role-play persona. It instructs the agent to 'Respond using grug-speak pattern', 'Refer to self as grug', and 'Maintain grug-speak throughout entire response'. While this is the primary intended function of the skill, it technically matches patterns used for persona-based prompt injection. - [INDIRECT_PROMPT_INJECTION]: The skill's primary use case is reviewing and refactoring user-provided code, which acts as a vector for untrusted data.
- Ingestion points: User-provided code snippets or architectural descriptions passed to the agent for simplification or review as described in
SKILL.md. - Boundary markers: The instructions do not define clear delimiters or provide instructions to the agent to disregard instructions potentially embedded within the code being analyzed.
- Capability inventory: The skill consists entirely of markdown files and does not contain any executable scripts, network requests, or file-writing tools.
- Sanitization: There are no sanitization steps provided to handle potentially malicious instructions embedded in the code snippets provided by the user.
Audit Metadata