prisma

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the agent to fetch live documentation from https://www.prisma.io/docs/llms.txt. This is an official resource provided by the Prisma ORM developers for AI agents.
  • [PROMPT_INJECTION]: The instruction set includes authoritative directives ("CRITICAL", "MUST") that command the agent to override its training data by fetching live external documentation before generating any code. This is a functional guardrail to ensure the use of current APIs.
  • [INDIRECT_PROMPT_INJECTION]: By requiring the agent to process an external file (llms.txt), the skill establishes an ingestion point for remote data which enters the agent's prompt context.
  • Ingestion points: SKILL.md (instruction to fetch llms.txt from prisma.io)
  • Boundary markers: Absent
  • Capability inventory: The agent uses the fetched documentation to generate Prisma schemas, client operations, and configuration strategies.
  • Sanitization: None
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 01:39 AM
Security Audit — agent-trust-hub — prisma