space-concept-board

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to ingest user-provided data, including text input and external files (e.g., existing moodboards or item lists), which are then processed by a team of agents. This creates a surface for indirect prompt injection where malicious instructions embedded in these files could influence agent behavior.
  • Ingestion points: Phase 1 explicitly extracts space information and constraints from user input and copies user-provided files directly into the _workspace/ directory for use in subsequent phases.
  • Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate between the agent's instructions and the content of user-provided data.
  • Capability inventory: The agent team has the capability to perform file read/write operations within the workspace, create and update tasks, and communicate via SendMessage to coordinate work.
  • Sanitization: There is no evidence of sanitization or validation performed on the user-provided files before they are utilized by the analysts and designers.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 05:20 AM
Security Audit — agent-trust-hub — space-concept-board