project-validator

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from project files, creating a potential surface for indirect prompt injection.
  • Ingestion points: The agent is instructed to read package.json and granite.config.ts from the local project directory (SKILL.md).
  • Boundary markers: The instructions do not specify the use of specific delimiters or "ignore embedded instructions" warnings when processing these files.
  • Capability inventory: The agent has file-reading capabilities and the instructions reference the existence of local validation scripts such as scripts/validate-all.ts (references/validation-policies.md).
  • Sanitization: No sanitization or strict schema validation of the raw file content is performed by the agent before the content is incorporated into the prompt context.
  • [EXTERNAL_DOWNLOADS]: The skill references several external packages and a repository for project setup and validation purposes.
  • References the official Apps-in-Toss Unity SDK repository for installation: https://github.com/toss/apps-in-toss-unity-sdk.git.
  • Suggests the installation of official framework and design system packages from the @apps-in-toss, @granite-js, and @toss scopes via standard package managers.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:18 AM
Security Audit — agent-trust-hub — project-validator