stitch-mcp-get-project
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes metadata from the external Stitch service, including potentially user-controlled fields such as project titles and design theme settings. This interpolation of external data into the agent's context creates a surface for indirect prompt injection if malicious instructions are embedded in the project metadata. \n- Ingestion points: Detailed project metadata (title, theme, screen instances) returned by the
get_projecttool (SKILL.md). \n- Boundary markers: No explicit delimiters or boundary instructions are defined for processing tool output. \n- Capability inventory: Primarily limited to read-only metadata discovery; instructions explicitly prohibit the skill from creating, modifying, or deleting resources. \n- Sanitization: No specific sanitization or validation logic for external project metadata is mentioned.
Audit Metadata