context-bridge
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- Prompt Injection (LOW): The skill facilitates an indirect prompt injection vector through its core context-sharing mechanism.
- Ingestion points: The agent is instructed to scan external project files (lore, mechanics, and asset descriptions) in the
World-BuildingandUnitydirectories (File:SKILL.md). - Boundary markers: The shared context file (
PROJECTS_SHARED_CONTEXT.md) does not utilize delimiters or specific instructions to the agent to ignore embedded commands within the summarized content. - Capability inventory: The skill uses
scripts/update_context.pyto write summarized data to the file system. This file is intended to be read by other sessions or projects to inform behavior. - Sanitization: The
update_context.pyscript takes the--contentargument and writes it directly to the shared file using regex replacement without any sanitization, validation, or escaping of the input string.
Audit Metadata