multica-projects-and-resources
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and act upon durable project context, which can include untrusted data from project descriptions or resource metadata.\n
- Ingestion points: Project descriptions and resource attributes (GitHub URLs, local directory paths) are fetched from the server and provided to the agent via task briefs and the
.multica/project/resources.jsonfile.\n - Boundary markers: Durable context is explicitly injected under the
## Project Contextheader in the agent's instructions, providing a structural boundary.\n - Capability inventory: The agent's capabilities are limited to the
multicaCLI tool as defined in theallowed-toolsfrontmatter.\n - Sanitization: The skill does not define specific sanitization routines for project descriptions or resource labels before they are integrated into the agent's context.
Audit Metadata