eigen-restaking
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (LOW): The skill instructions suggest using the
exectool to runcurlcommands targeting the EigenExplorer API. These commands are specific to data retrieval and do not involve executing untrusted remote scripts or shell pipes. - [DATA_EXFILTRATION] (LOW): The skill performs network requests to
api.eigenexplorer.comandapi-holesky.eigenexplorer.com. These are non-whitelisted domains, though the activity is strictly aligned with the skill's stated purpose and handles credentials via environment variables. - [PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection from processing external API data. Ingestion points: API responses from
api.eigenexplorer.comvia scripts/eigen-api.js and manual curl commands. Boundary markers: Absent; no instructions provided to the agent to ignore instructions within the retrieved data. Capability inventory: Includes subprocess execution viacurland potential file writes if the agent decides to cache data. Sanitization: None identified in the provided code.
Audit Metadata