eve-auth-and-secrets
Warn
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructions specify access to sensitive local file paths for authentication and secret management.
- The
eve secrets import .envcommand allows the agent to ingest environment files which typically contain sensitive credentials and API keys. - Documentation references the synchronization and management of authentication files located at
~/.codex/auth.jsonand~/.code/auth.json. - Instructions for authentication and inviting users involve reading SSH public keys from the
~/.ssh/directory. - [EXTERNAL_DOWNLOADS]: The skill references external Node.js packages for application integration.
- Integration guidelines require the installation of
@eve-horizon/authand@eve-horizon/auth-reactvia npm. - [COMMAND_EXECUTION]: The skill provides comprehensive instructions for using the
eveCLI tool to perform administrative and developmental tasks. - Commands include authentication (
eve auth login), secret manipulation (eve secrets set), and organization management (eve admin invite). - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where external or untrusted data could potentially influence agent behavior.
- Ingestion points: The skill ingests data from local
.envfiles, user-supplied input for invitations (emails, names), and responses from external API endpoints. - Boundary markers: The instructions do not define explicit delimiters to separate untrusted data from agent instructions.
- Capability inventory: The skill facilitates file reading, network communication with the Eve API, and secret synchronization, providing a significant capability set if injection occurs.
- Sanitization: The skill includes manifest validation commands (
eve manifest validate) to check for required secrets.
Audit Metadata