ios-design-mockup
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external files (PRDs, user stories) to generate a visual HTML mockup. This creates a surface where a malicious specification could attempt to influence the agent's behavior or inject malicious scripts into the generated output. Ingestion points:
SKILL.md(Step 1). Capability inventory: File writing inSKILL.md(Step 4). Boundary markers: Absent. Sanitization: Not specified for the injected content. - [DATA_EXFILTRATION]: The skill is designed to read from user-provided file paths to extract specifications. While intended for project documentation, this capability could be abused if an attacker tricks the agent into reading sensitive files (such as credentials, SSH keys, or environment files) by providing their paths as the 'spec-path'. Evidence:
SKILL.mdStep 1 explicitly instructs the agent to read from paths mentioned by the user.
Audit Metadata