grafana-dashboards
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could contain malicious instructions.
- Ingestion points: Dashboard JSON configurations are retrieved and processed in the workflow (SKILL.md, Workflow steps 3 and 5) via the
homelab_grafanaSnapshottool. - Boundary markers: No explicit delimiters or instructions to ignore embedded instructions within the processed JSON data are present.
- Capability inventory: The skill utilizes tools with side effects, including service restarts (
homelab_serviceRestart), service log inspection (homelab_serviceLogs), and querying data sources (homelab_prometheusQuery). - Sanitization: No sanitization or validation logic is specified for the JSON content before it is modified or re-imported.
- [CREDENTIALS_UNSAFE]: The skill documentation hardcodes well-known default credentials (
admin/admin) for the Grafana service. While the documentation correctly advises the user to change these, the inclusion of literal credentials in instructions is a security best-practice violation.
Audit Metadata