eve
Warn
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill directs the agent to install the "eve" package via "npm install eve" or "npx eve init". On the public NPM registry, the "eve" package is a legacy event library and does not match the "filesystem-first framework for durable backend AI agents" described in the skill. This discrepancy creates a risk where the agent may install and execute code that is unrelated to the intended task or potentially malicious.\n- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to treat external files bundled with the package as its primary source of truth.\n
- Ingestion points: The agent is directed to read "node_modules/eve/docs/README.md" and other documents in that directory.\n
- Boundary markers: None are provided; the skill explicitly tells the agent to disregard its own instructions in favor of the external documentation.\n
- Capability inventory: The agent's purpose involves creating and editing projects, which implies filesystem access and command execution that could be exploited by malicious documentation content.\n
- Sanitization: No validation of the external content is performed before the agent processes it.\n- [EXTERNAL_DOWNLOADS]: The skill promotes the installation of a third-party package from a public registry without version pinning or verification of the package content against the described functionality.
Audit Metadata