high-quality-vibe-coding
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous examples of shell commands for developers and agents to use for project setup, quality assurance, and verification.
- Evidence: Standard development commands like
npm install -D husky lint-staged,npx biome check --write,tsc --noEmit,eslint ., andvitest runare included to enforce code quality. - [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow that involves reading project files and screenshots, which represents a potential attack surface for indirect prompt injection from untrusted source code or visual data.
- Ingestion points: The skill instructs agents to analyze existing code patterns in the repository and verify UI changes through screenshots.
- Boundary markers: It recommends utilizing
CLAUDE.mdand.cursor/rulesto establish explicit instructions and constraints for the agent's behavior. - Capability inventory: The skill employs
Read,Write,Edit, andBashtools for file manipulation and command execution. - Sanitization: The methodology strictly mandates a 'Three Laws' approach requiring human verification (Diff Review), automated tests (TDD), and compiler-level checks (TypeScript strict mode) to ensure all AI-generated or influenced code is safe and correct.
Audit Metadata