env-manager
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The documentation provides Python patterns using
subprocess.runto interact with external command-line tools including Heroku, Railway, and 1Password (op). Each interaction is mapped to specific management tasks such as setting configuration variables.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from environment files (.env) and provides capabilities to execute commands and perform network requests based on that data.\n - Ingestion points: Environment files provided by the user for validation (e.g., .env, .env.local).\n
- Boundary markers: Standard parsing is assumed; no specific instructions are provided to the agent to disregard instructions found within values.\n
- Capability inventory: The skill snippets include file reading/writing, network operations via the
requestslibrary, and command execution usingsubprocess.\n - Sanitization: Patterns for masking secret values are included to prevent credential leakage in logs.\n- [EXTERNAL_DOWNLOADS]: The skill references and provides code for interacting with well-known cloud services and APIs.\n
- Evidence: Integration code is provided for Vercel (api.vercel.com), Railway (backboard.railway.app), and AWS Secrets Manager. These are recognized as well-known service providers.
Audit Metadata