steedos-builder6-modules
Pass
Audited by Gen Agent Trust Hub on May 6, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The plugin system documentation describes a mechanism for dynamic code loading. It specifies that NestJS modules are loaded from "dist/plugin.module.js" and Moleculer services are initialized via "MoleculerPluginService".
- [COMMAND_EXECUTION]: The skill details the use of "npm install --omit=dev --no-audit" for managing and updating system plugins based on environment variable configurations.
- [EXTERNAL_DOWNLOADS]: Documents the retrieval of external software components from NPM registries, including support for private registries using the "B6_PLUGIN_NPMRC" variable.
- [DATA_EXFILTRATION]: Outlines file management capabilities, including S3 storage integration and the generation of pre-signed URLs. It provides configuration templates for AWS S3 authentication using placeholders.
- [PROMPT_INJECTION]: The skill documents interfaces that ingest untrusted data, creating a surface for indirect prompt injection. Ingestion points: File upload endpoints ("/api/v6/files/:collectionName") and plugin package definitions ("B6_PLUGIN_PACKAGES"). Boundary markers: None specified in the documentation. Capability inventory: Includes dynamic code execution ("require"), package installation ("npm install"), and file storage operations (Local/S3). Sanitization: Not described in the current skill content.
Audit Metadata