mirrord-prev-env
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consistently recommends security best practices, such as using the least-privilege
mirrord-operator-ciClusterRole and gating preview starts on trusted pull requests. - [SAFE]: External references to the
metalbear-co/mirrord-previewGitHub Action andmetalbear/mirrord-operator-share-ingressHelm chart are official vendor resources and are documented neutrally. - [INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection as it processes user-provided configuration files to drive system behavior.
- Ingestion points: The skill ingests untrusted data from
mirrord.json,extra_config, and various CLI/Action inputs. - Boundary markers: The instructions explicitly warn the agent to treat these inputs as untrusted data rather than instructions and to avoid executing shell commands derived from their values.
- Capability inventory: The skill has the capability to create and manage Kubernetes resources (Pods, Services, CronJobs) and route network traffic via the mirrord Operator.
- Sanitization: The skill includes instructions to validate configurations and clearly advises against executing shell commands or fetching URLs found within the untrusted input data.
Audit Metadata