signoz-creating-dashboards
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill uses natural instructional language and clear constraints without attempting to override system safety guidelines or extract internal prompts. It includes explicit instructions to stay within the provided toolset.
- [DATA_EXFILTRATION]: No sensitive data access or external network operations were detected. The skill specifically warns against making raw HTTP calls or fetching content from external URLs like GitHub, insisting on using internal SigNoz MCP tools.
- [REMOTE_CODE_EXECUTION]: The skill does not perform remote code execution. It uses a defined set of MCP tools (e.g.,
signoz_create_dashboard,signoz_import_dashboard) to interact with the SigNoz API. It explicitly forbids shell execution. - [INDIRECT_PROMPT_INJECTION]: The skill processes dashboard configurations and templates retrieved from the SigNoz server. While this presents an ingestion surface for untrusted data, the risk is mitigated by the skill's instruction to treat the JSON as data structures for API calls rather than executable instructions. Boundary markers are absent, but capabilities are restricted to observability operations.
- [DYNAMIC_EXECUTION]: The skill builds JSON objects for dashboards based on local MCP documentation (
signoz://dashboard/*). This is a legitimate configuration task and does not involve executing dynamically generated code or unsafe deserialization.
Audit Metadata