prisma-workflow
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- COMMAND_EXECUTION (MEDIUM): The agent is instructed to add commands to 'package.json' and execute 'npm run db:test' during the verification phase. Automated shell command execution increases the risk of command injection.
- REMOTE_CODE_EXECUTION (MEDIUM): In Phase 4, the agent must generate 'scripts/test-database.ts' and execute it. This 'write-then-run' pattern is dangerous if the content of the script is influenced by external, untrusted data like database model definitions or schemas.
- EXTERNAL_DOWNLOADS (LOW): The skill uses 'npx' and installs '@prisma/adapter-pg' and other dependencies. While these are from the trusted npm registry, they represent runtime dependency fetching.
- INDIRECT_PROMPT_INJECTION (MEDIUM): As a database workflow tool, the skill likely processes untrusted schemas or configuration data. Because the skill has the capability to write and execute code based on these inputs, it is vulnerable to indirect prompt injection attacks where malicious data triggers harmful script generation.
Audit Metadata