skills/somtougeh/dotfiles/prisma-cli/Gen Agent Trust Hub

prisma-cli

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: Documentation for 'prisma db execute' in 'references/db-execute.md' describes the ability to execute arbitrary SQL commands from files or standard input, which allows for direct database manipulation.\n- [COMMAND_EXECUTION]: Documentation for destructive commands in 'references/migrate-reset.md' and 'references/db-push.md' covers resetting the database and forcing schema changes, which are high-impact operations that can lead to total data loss.\n- [REMOTE_CODE_EXECUTION]: Documentation for 'prisma db seed' in 'references/db-seed.md' details how to execute external scripts defined in the 'prisma.config.ts' file using runtimes like 'tsx', 'ts-node', or 'node'.\n- [DATA_EXFILTRATION]: Documentation for 'prisma studio' in 'references/studio.md' and 'prisma db pull' in 'references/db-pull.md' describes tools that provide visibility into database contents and structure, representing a surface for data exposure.\n- [PROMPT_INJECTION]: An indirect prompt injection surface is present via the 'prisma init --prompt' feature in 'SKILL.md'. Ingestion points: Natural language input used for schema generation and deployment. Boundary markers: Absent. Capability inventory: Automated schema generation and database deployment. Sanitization: None documented within the CLI reference instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 07:09 AM
Security Audit — agent-trust-hub — prisma-cli