cloud-manage-project
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is authored by Elastic and communicates exclusively with official Elastic API endpoints (api.elastic-cloud.com).
- [COMMAND_EXECUTION]: The skill uses 'eval' on output from a local Python script to set environment variables. The script uses 'shlex.quote' to sanitize all values, mitigating command injection risks.
- [DATA_EXFILTRATION]: Credential management is performed via '.env' and '.elastic-credentials' files. The skill includes robust guardrails that explicitly forbid the agent from displaying, echoing, or logging these secrets in the chat or thinking traces.
- [PROMPT_INJECTION]: The skill ingests project metadata from Elastic API responses (Ingestion: manage-project.py). Boundary markers are provided via instructional context. Capability inventory includes network access, file writing, and environment variable manipulation. Sanitization is implemented via shlex.quote. The risk of indirect prompt injection is minimal due to the trusted nature of the data source.
Audit Metadata