start-building-for-startups
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to scan project files (e.g., package.json, configuration files) and read user-provided documentation or URLs to gather context. This creates a surface for indirect prompt injection if the codebase or external documents contain malicious instructions designed to subvert the agent's logic during the implementation phase.
- Ingestion points: Local project structure, configuration files (Dockerfile, IaC, package manifests), and user-supplied external links or design files.
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions to isolate project data from operational logic.
- Capability inventory: The agent can write implementation code directly into the project and provides commands for environment configuration.
- Sanitization: No explicit sanitization or validation of the ingested codebase content is described.
- [COMMAND_EXECUTION]: The skill provides the agent with instructions to guide the user through setting up their environment, including the installation of the AWS CLI, running
aws configurefor credentials, and verifying identity viaaws sts get-caller-identity. - [SAFE]: The skill identifies itself as a product of 'awslabs' and focuses exclusively on AWS-native services and architectural patterns. All external file references (such as to the knowledge base or offers index) are relative paths to sibling skills, representing a standard suite of development tools from the same vendor.
Audit Metadata