interface-design
Pass
Audited by Gen Agent Trust Hub on Jul 5, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious code or scripts were found in the skill files. The content consists entirely of instructional markdown.
- [NO_CODE]: The skill does not include any executable scripts, binaries, or automated installation steps, significantly reducing the attack surface.
- [PROMPT_INJECTION]: The skill establishes an indirect prompt injection surface by instructing the agent to read and write design patterns to a local file (
.interface-design/system.md). This allows for context persistence across sessions but could be exploited if an attacker modifies project files. - Ingestion points: File read operations for
.interface-design/system.mdreferenced inSKILL.mdandreferences/validation.md. - Boundary markers: Absent; there are no specific instructions to treat the file content as data rather than instructions.
- Capability inventory: The skill instructions involve the agent writing to the local file system using its native tools.
- Sanitization: No sanitization or validation of the data stored in the system file is required by the instructions.
Audit Metadata