The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials or sensitive data exposure patterns were found. The configuration schema in schemas/config-schema.json correctly defines fields for api_key as configuration parameters without providing actual values.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): No remote scripts, package installations, or piped shell commands (e.g., curl | bash) are present in the templates or documentation.
[Indirect Prompt Injection] (SAFE): The architecture described in references/reference.md defines a workflow for processing specifications (SPECs) into documentation. This is a common ingestion surface but no malicious patterns were detected in the templates themselves. 1. Ingestion points: spec_data passed to generate_documentation_from_spec in references/reference.md. 2. Boundary markers: Not explicitly defined in the provided templates. 3. Capability inventory: File system write operations for documentation generation and project initialization. 4. Sanitization: Implementation details are not provided in these static templates.
[Dynamic Execution] (SAFE): The skill documentation and schema mention support for template engines like Jinja2 and Handlebars. In the context of these configuration files, these are descriptive options and do not constitute a dynamic execution vulnerability.

moai-workflow-project